Publications
OUR RESEARCH
Scientific Publications
Here you can find the comprehensive list of publications from the members of the Research Center on Computer Vision and eXtended Reality (xRAI).
Use the tag cloud to filter papers based on specific research topics, or use the menus to filter by year, type of publication, or authors.
For each paper, you have the option to view additional details such as the Abstract, Links, and BibTex record.
Research is formalized curiosity. It is poking and prying with a purpose
Zora Neale Hurston
2020
Galteri, Leonardo; Seidenari, Lorenzo; Uricchio, Tiberio; Bertini, Marco; Bimbo, Alberto Del
Preserving low-quality video through deep learning Proceedings Article
In: IOP Conference Series: Materials Science and Engineering, pp. 012068, IOP Publishing, 2020, ISBN: 1757-899X, (tex.copyright: All rights reserved).
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Compression Artifact Removal, Computer Vision and Pattern Recognition, Digital Archiving, Generative Adversarial Networks, Image Restoration, Video Streaming
@inproceedings{galteriPreservingLowqualityVideo2020,
title = {Preserving low-quality video through deep learning},
author = {Leonardo Galteri and Lorenzo Seidenari and Tiberio Uricchio and Marco Bertini and Alberto Del Bimbo},
url = {https://iopscience.iop.org/article/10.1088/1757-899X/949/1/012068/pdf},
doi = {10.1088/1757-899X/949/1/012068},
isbn = {1757-899X},
year = {2020},
date = {2020-01-01},
booktitle = {IOP Conference Series: Materials Science and Engineering},
volume = {949},
pages = {012068},
publisher = {IOP Publishing},
abstract = {Lossy video stream compression is performed to reduce the bandwidth and storage requirements. Moreover also image compression is a need that arises in many circumstances.It is often the case that older archive are stored at low resolution and with a compression rate suitable for the technology available at the time the video was created. Unfortunately, lossy compression algorithms cause artifact. Such artifacts, usually damage higher frequency details also adding noise or novel image patterns. There are several issues with this phenomenon. Low-quality images can be less pleasant to persons. Object detectors algorithms may have their performance reduced. As a result, given a perturbed version of it, we aim at removing such artifacts to recover the original image. To obtain that, one should reverse the compression process through a complicated non-linear image transformation. We propose a deep neural network able to improve image quality. We show that this model can be optimized either traditionally, directly optimizing an image similarity loss (SSIM), or using a generative adversarial approach (GAN). Our restored images have more photorealistic details with respect to traditional image enhancement networks. Our training procedure based on sub-patches is novel. Moreover, we propose novel testing protocol to evaluate restored images quantitatively. Differently from previously proposed approaches we are able to remove artifacts generated at any quality by inferring the image quality directly from data. Human evaluation and quantitative experiments in object detection show that our GAN generates images with finer consistent details and these details make a difference both for machines and humans.},
note = {tex.copyright: All rights reserved},
keywords = {Artificial Intelligence, Compression Artifact Removal, Computer Vision and Pattern Recognition, Digital Archiving, Generative Adversarial Networks, Image Restoration, Video Streaming},
pubstate = {published},
tppubtype = {inproceedings}
}
Lossy video stream compression is performed to reduce the bandwidth and storage requirements. Moreover also image compression is a need that arises in many circumstances.It is often the case that older archive are stored at low resolution and with a compression rate suitable for the technology available at the time the video was created. Unfortunately, lossy compression algorithms cause artifact. Such artifacts, usually damage higher frequency details also adding noise or novel image patterns. There are several issues with this phenomenon. Low-quality images can be less pleasant to persons. Object detectors algorithms may have their performance reduced. As a result, given a perturbed version of it, we aim at removing such artifacts to recover the original image. To obtain that, one should reverse the compression process through a complicated non-linear image transformation. We propose a deep neural network able to improve image quality. We show that this model can be optimized either traditionally, directly optimizing an image similarity loss (SSIM), or using a generative adversarial approach (GAN). Our restored images have more photorealistic details with respect to traditional image enhancement networks. Our training procedure based on sub-patches is novel. Moreover, we propose novel testing protocol to evaluate restored images quantitatively. Differently from previously proposed approaches we are able to remove artifacts generated at any quality by inferring the image quality directly from data. Human evaluation and quantitative experiments in object detection show that our GAN generates images with finer consistent details and these details make a difference both for machines and humans.
Tomazzoli, Claudio; Scannapieco, Simone; Cristani, Matteo
Internet of Things and Artificial Intelligence Enable Energy Efficiency Journal Article
In: Journal of Ambient Intelligence and Humanized Computing, 2020.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Energy Efficiency, Intelligent Systems, Machine Learning
@article{tomazzoli_internet_2020,
title = {Internet of Things and Artificial Intelligence Enable Energy Efficiency},
author = {Claudio Tomazzoli and Simone Scannapieco and Matteo Cristani},
url = {https://doi.org/10.1007/s12652-020-02151-3},
doi = {10.1007/s12652-020-02151-3},
year = {2020},
date = {2020-01-01},
journal = {Journal of Ambient Intelligence and Humanized Computing},
abstract = {In smart environments, there is an increasing demand for scalable and autonomous management systems. In this regard, energy efficiency hands out challenging aspects, for both home and business usages. Scalability in energy management systems is particularly difficult in those industry sector where power consumption of branches located in remote areas need to be monitored. Being autonomous requires that behavioural rules are automatically extracted from consumption data and applied to the system. Best practices for the specific energy configuration should be devised to achieve optimal energy efficiency. Best practices should also be revised and applied without human intervention against topology changes. In this paper, the Internet of Things paradigm and machine learning techniques are exploited to (1) define a novel system architecture for centralised energy efficiency in distributed sub-networks of electric appliances, (2) extract behavioural rules, identify best practices and detect device types. A system architecture tailored for autonomous energy efficiency has interesting applications in smart industry—where energy managers may effortlessly monitor and optimally setup a large number of sparse divisions—and smart home—where impaired people may avoid energy waste through an autonomous system that can be employed by the users as a delegate for decision making.},
keywords = {Artificial Intelligence, Energy Efficiency, Intelligent Systems, Machine Learning},
pubstate = {published},
tppubtype = {article}
}
In smart environments, there is an increasing demand for scalable and autonomous management systems. In this regard, energy efficiency hands out challenging aspects, for both home and business usages. Scalability in energy management systems is particularly difficult in those industry sector where power consumption of branches located in remote areas need to be monitored. Being autonomous requires that behavioural rules are automatically extracted from consumption data and applied to the system. Best practices for the specific energy configuration should be devised to achieve optimal energy efficiency. Best practices should also be revised and applied without human intervention against topology changes. In this paper, the Internet of Things paradigm and machine learning techniques are exploited to (1) define a novel system architecture for centralised energy efficiency in distributed sub-networks of electric appliances, (2) extract behavioural rules, identify best practices and detect device types. A system architecture tailored for autonomous energy efficiency has interesting applications in smart industry—where energy managers may effortlessly monitor and optimally setup a large number of sparse divisions—and smart home—where impaired people may avoid energy waste through an autonomous system that can be employed by the users as a delegate for decision making.
Caggianese, Giuseppe; Pietro, Giuseppe De; Esposito, Massimo; Gallo, Luigi; Minutolo, Aniello; Neroni, Pietro
Discovering Leonardo with artificial intelligence and holograms: A user study Journal Article
In: Pattern Recognition Letters, vol. 131, pp. 361–367, 2020, ISSN: 0167-8655.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Conversational systems, Cultural Heritage, Holograms, Human Computer Interaction, Touchless interaction, User study
@article{caggianeseDiscoveringLeonardoArtificial2020,
title = {Discovering Leonardo with artificial intelligence and holograms: A user study},
author = {Giuseppe Caggianese and Giuseppe De Pietro and Massimo Esposito and Luigi Gallo and Aniello Minutolo and Pietro Neroni},
url = {https://www.sciencedirect.com/science/article/pii/S0167865520300039},
doi = {https://doi.org/10.1016/j.patrec.2020.01.006},
issn = {0167-8655},
year = {2020},
date = {2020-01-01},
journal = {Pattern Recognition Letters},
volume = {131},
pages = {361–367},
abstract = {Cutting-edge visualization and interaction technologies are increasingly used in museum exhibitions, providing novel ways to engage visitors and enhance their cultural experience. Existing applications are commonly built upon a single technology, focusing on visualization, motion or verbal interaction (e.g., high-resolution projections, gesture interfaces, chatbots). This aspect limits their potential, since museums are highly heterogeneous in terms of visitors profiles and interests, requiring multi-channel, customizable interaction modalities. To this aim, this work describes and evaluates an artificial intelligence powered, interactive holographic stand aimed at describing Leonardo Da Vinci’s art. This system provides the users with accurate 3D representations of Leonardo’s machines, which can be interactively manipulated through a touchless user interface. It is also able to dialog with the users in natural language about Leonardo’s art, while keeping the context of conversation and interactions. Furthermore, the results of a large user study, carried out during art and tech exhibitions, are presented and discussed. The goal was to assess how users of different ages and interests perceive, understand and explore cultural objects when holograms and artificial intelligence are used as instruments of knowledge and analysis.},
keywords = {Artificial Intelligence, Conversational systems, Cultural Heritage, Holograms, Human Computer Interaction, Touchless interaction, User study},
pubstate = {published},
tppubtype = {article}
}
Cutting-edge visualization and interaction technologies are increasingly used in museum exhibitions, providing novel ways to engage visitors and enhance their cultural experience. Existing applications are commonly built upon a single technology, focusing on visualization, motion or verbal interaction (e.g., high-resolution projections, gesture interfaces, chatbots). This aspect limits their potential, since museums are highly heterogeneous in terms of visitors profiles and interests, requiring multi-channel, customizable interaction modalities. To this aim, this work describes and evaluates an artificial intelligence powered, interactive holographic stand aimed at describing Leonardo Da Vinci’s art. This system provides the users with accurate 3D representations of Leonardo’s machines, which can be interactively manipulated through a touchless user interface. It is also able to dialog with the users in natural language about Leonardo’s art, while keeping the context of conversation and interactions. Furthermore, the results of a large user study, carried out during art and tech exhibitions, are presented and discussed. The goal was to assess how users of different ages and interests perceive, understand and explore cultural objects when holograms and artificial intelligence are used as instruments of knowledge and analysis.
Canonico, Roberto; Cozzolino, Giovanni; Ferraro, Antonino; Moscato, Vincenzo; Picariello, Antonio; Sorrentino, Fabio Raimondo; Sperlì, Giancarlo
A Smart ChatBot for Specialist Domains Book Section
In: Barolli, Leonard; Amato, Flora; Moscato, Francesco; Enokido, Tomoya; Takizawa, Makoto (Ed.): Web, Artificial Intelligence and Network Applications, vol. 1150, pp. 1003–1010, Springer International Publishing, Cham, 2020, ISBN: 978-3-030-44037-4 978-3-030-44038-1, (Series Title: Advances in Intelligent Systems and Computing).
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Chatbot, Cultural Heritage
@incollection{canonicoSmartChatBotSpecialist2020,
title = {A Smart ChatBot for Specialist Domains},
author = {Roberto Canonico and Giovanni Cozzolino and Antonino Ferraro and Vincenzo Moscato and Antonio Picariello and Fabio Raimondo Sorrentino and Giancarlo Sperlì},
editor = {Leonard Barolli and Flora Amato and Francesco Moscato and Tomoya Enokido and Makoto Takizawa},
url = {https://link.springer.com/10.1007/978-3-030-44038-1_92},
doi = {10.1007/978-3-030-44038-1_92},
isbn = {978-3-030-44037-4 978-3-030-44038-1},
year = {2020},
date = {2020-01-01},
urldate = {2024-07-12},
booktitle = {Web, Artificial Intelligence and Network Applications},
volume = {1150},
pages = {1003–1010},
publisher = {Springer International Publishing},
address = {Cham},
abstract = {Artificial Intelligence application are becoming increasingly complex and sophisticated in order to shorten the gap between the users and digital systems. Nowadays, researchers and companies are making a big effort to develop new forms of interactions with a growing number and kinds of electronic devices, with a particular attention on the conversational interface, spoken or written. A chatbot is software that simulates and processes human conversations, allowing users to interact with digital devices as if they were communicating with a real person. In this paper we introduce a Chatbot for the cultural heritage domain [1], able to respond to a simple single-line query or to suggest users, as a digital assistants, by providing increasing levels of customisation through the collection and the processing of users’ conversations.},
note = {Series Title: Advances in Intelligent Systems and Computing},
keywords = {Artificial Intelligence, Chatbot, Cultural Heritage},
pubstate = {published},
tppubtype = {incollection}
}
Artificial Intelligence application are becoming increasingly complex and sophisticated in order to shorten the gap between the users and digital systems. Nowadays, researchers and companies are making a big effort to develop new forms of interactions with a growing number and kinds of electronic devices, with a particular attention on the conversational interface, spoken or written. A chatbot is software that simulates and processes human conversations, allowing users to interact with digital devices as if they were communicating with a real person. In this paper we introduce a Chatbot for the cultural heritage domain [1], able to respond to a simple single-line query or to suggest users, as a digital assistants, by providing increasing levels of customisation through the collection and the processing of users’ conversations.
Generosi, Andrea; Ceccacci, Silvia; Faggiano, Samuele; Giraldi, Luca; Mengoni, Maura
A Toolkit for the Automatic Analysis of Human Behavior in HCI Applications in the Wild Journal Article
In: Advances in Science, Technology and Engineering Systems Journal, vol. 5, no. 6, pp. 185–192, 2020, ISSN: 24156698, 24156698.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Deep Learning, Human Computer Interaction
@article{generosi_toolkit_2020,
title = {A Toolkit for the Automatic Analysis of Human Behavior in HCI Applications in the Wild},
author = {Andrea Generosi and Silvia Ceccacci and Samuele Faggiano and Luca Giraldi and Maura Mengoni},
url = {https://astesj.com/v05/i06/p22/},
doi = {10.25046/aj050622},
issn = {24156698, 24156698},
year = {2020},
date = {2020-01-01},
urldate = {2024-12-28},
journal = {Advances in Science, Technology and Engineering Systems Journal},
volume = {5},
number = {6},
pages = {185–192},
abstract = {Nowadays, smartphones and laptops equipped with cameras have become an integral part of our daily lives. The pervasive use of cameras enables the collection of an enormous amount of data, which can be easily extracted through video images processing. This opens up the possibility of using technologies that until now had been restricted to laboratories, such as eye-tracking and emotion analysis systems, to analyze users' behavior in the wild, during the interaction with websites. In this context, this paper introduces a toolkit that takes advantage of deep learning algorithms to monitor user's behavior and emotions, through the acquisition of facial expression and eye gaze from the video captured by the webcam of the device used to navigate the web, in compliance with the EU General data protection regulation (GDPR). Collected data are potentially useful to support user experience assessment of web-based applications in the wild and to improve the effectiveness of e-commerce recommendation systems.},
keywords = {Artificial Intelligence, Deep Learning, Human Computer Interaction},
pubstate = {published},
tppubtype = {article}
}
Nowadays, smartphones and laptops equipped with cameras have become an integral part of our daily lives. The pervasive use of cameras enables the collection of an enormous amount of data, which can be easily extracted through video images processing. This opens up the possibility of using technologies that until now had been restricted to laboratories, such as eye-tracking and emotion analysis systems, to analyze users' behavior in the wild, during the interaction with websites. In this context, this paper introduces a toolkit that takes advantage of deep learning algorithms to monitor user's behavior and emotions, through the acquisition of facial expression and eye gaze from the video captured by the webcam of the device used to navigate the web, in compliance with the EU General data protection regulation (GDPR). Collected data are potentially useful to support user experience assessment of web-based applications in the wild and to improve the effectiveness of e-commerce recommendation systems.
Ceccacci, Silvia; Mengoni, Maura; Generosi, Andrea; Giraldi, Luca; Carbonara, Giuseppe; Castellano, Andrea; Montanari, Roberto
A Preliminary Investigation Towards the Application of Facial Expression Analysis to Enable an Emotion-Aware Car Interface Proceedings Article
In: Antona, Margherita; Stephanidis, Constantine (Ed.): Universal Access in Human-Computer Interaction. Applications and Practice, pp. 504–517, Springer International Publishing, Cham, 2020, ISBN: 978-3-030-49107-9 978-3-030-49108-6.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Emotion Recognition, Human Computer Interaction
@inproceedings{antona_preliminary_2020,
title = {A Preliminary Investigation Towards the Application of Facial Expression Analysis to Enable an Emotion-Aware Car Interface},
author = {Silvia Ceccacci and Maura Mengoni and Andrea Generosi and Luca Giraldi and Giuseppe Carbonara and Andrea Castellano and Roberto Montanari},
editor = {Margherita Antona and Constantine Stephanidis},
url = {https://link.springer.com/10.1007/978-3-030-49108-6_36},
doi = {10.1007/978-3-030-49108-6_36},
isbn = {978-3-030-49107-9 978-3-030-49108-6},
year = {2020},
date = {2020-01-01},
urldate = {2024-12-28},
booktitle = {Universal Access in Human-Computer Interaction. Applications and Practice},
volume = {12189},
pages = {504–517},
publisher = {Springer International Publishing},
address = {Cham},
abstract = {The paper describes the conceptual model of an emotion-aware car interface able to: map both the driver’s cognitive and emotional states with the vehicle dynamics; adapt the level of automation or support the decision-making process if emotions negatively affecting the driving performance are detected; ensure emotion regulation and provide a unique user experience creating a more engaging atmosphere (e.g. music, LED lighting) in the car cabin. To enable emotion detection, it implements a low-cost emotion recognition able to recognize Ekman’s universal emotions by analyzing the driver’s facial expression from stream video. A preliminary test was conducted in order to determine the effectiveness of the proposed emotion recognition system in a driving context. Results evidenced that the proposed system is capable to correctly qualify the drivers’ emotion in a driving simulation context.},
keywords = {Artificial Intelligence, Emotion Recognition, Human Computer Interaction},
pubstate = {published},
tppubtype = {inproceedings}
}
The paper describes the conceptual model of an emotion-aware car interface able to: map both the driver’s cognitive and emotional states with the vehicle dynamics; adapt the level of automation or support the decision-making process if emotions negatively affecting the driving performance are detected; ensure emotion regulation and provide a unique user experience creating a more engaging atmosphere (e.g. music, LED lighting) in the car cabin. To enable emotion detection, it implements a low-cost emotion recognition able to recognize Ekman’s universal emotions by analyzing the driver’s facial expression from stream video. A preliminary test was conducted in order to determine the effectiveness of the proposed emotion recognition system in a driving context. Results evidenced that the proposed system is capable to correctly qualify the drivers’ emotion in a driving simulation context.
Canonico, Roberto; Cozzolino, Giovanni; Ferraro, Antonino; Moscato, Vincenzo; Picariello, Antonio; Sorrentino, Fabio Raimondo; Sperlì, Giancarlo
A Smart ChatBot for Specialist Domains Book Section
In: Barolli, Leonard; Amato, Flora; Moscato, Francesco; Enokido, Tomoya; Takizawa, Makoto (Ed.): Web, Artificial Intelligence and Network Applications, vol. 1150, pp. 1003–1010, Springer International Publishing, Cham, 2020, ISBN: 978-3-030-44037-4 978-3-030-44038-1, (Series Title: Advances in Intelligent Systems and Computing).
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Chatbot, Cultural Heritage
@incollection{barolli_smart_2020,
title = {A Smart ChatBot for Specialist Domains},
author = {Roberto Canonico and Giovanni Cozzolino and Antonino Ferraro and Vincenzo Moscato and Antonio Picariello and Fabio Raimondo Sorrentino and Giancarlo Sperlì},
editor = {Leonard Barolli and Flora Amato and Francesco Moscato and Tomoya Enokido and Makoto Takizawa},
url = {https://link.springer.com/10.1007/978-3-030-44038-1_92},
doi = {10.1007/978-3-030-44038-1_92},
isbn = {978-3-030-44037-4 978-3-030-44038-1},
year = {2020},
date = {2020-01-01},
urldate = {2024-07-12},
booktitle = {Web, Artificial Intelligence and Network Applications},
volume = {1150},
pages = {1003–1010},
publisher = {Springer International Publishing},
address = {Cham},
abstract = {Artificial Intelligence application are becoming increasingly complex and sophisticated in order to shorten the gap between the users and digital systems. Nowadays, researchers and companies are making a big effort to develop new forms of interactions with a growing number and kinds of electronic devices, with a particular attention on the conversational interface, spoken or written. A chatbot is software that simulates and processes human conversations, allowing users to interact with digital devices as if they were communicating with a real person. In this paper we introduce a Chatbot for the cultural heritage domain [1], able to respond to a simple single-line query or to suggest users, as a digital assistants, by providing increasing levels of customisation through the collection and the processing of users’ conversations.},
note = {Series Title: Advances in Intelligent Systems and Computing},
keywords = {Artificial Intelligence, Chatbot, Cultural Heritage},
pubstate = {published},
tppubtype = {incollection}
}
Artificial Intelligence application are becoming increasingly complex and sophisticated in order to shorten the gap between the users and digital systems. Nowadays, researchers and companies are making a big effort to develop new forms of interactions with a growing number and kinds of electronic devices, with a particular attention on the conversational interface, spoken or written. A chatbot is software that simulates and processes human conversations, allowing users to interact with digital devices as if they were communicating with a real person. In this paper we introduce a Chatbot for the cultural heritage domain [1], able to respond to a simple single-line query or to suggest users, as a digital assistants, by providing increasing levels of customisation through the collection and the processing of users’ conversations.
Caggianese, Giuseppe; Pietro, Giuseppe De; Esposito, Massimo; Gallo, Luigi; Minutolo, Aniello; Neroni, Pietro
Discovering Leonardo with artificial intelligence and holograms: A user study Journal Article
In: Pattern Recognition Letters, vol. 131, pp. 361–367, 2020, ISSN: 0167-8655.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Conversational systems, Cultural Heritage, Holograms, Human Computer Interaction, Touchless interaction, User study
@article{caggianese_discovering_2020,
title = {Discovering Leonardo with artificial intelligence and holograms: A user study},
author = {Giuseppe Caggianese and Giuseppe De Pietro and Massimo Esposito and Luigi Gallo and Aniello Minutolo and Pietro Neroni},
url = {https://www.sciencedirect.com/science/article/pii/S0167865520300039},
doi = {https://doi.org/10.1016/j.patrec.2020.01.006},
issn = {0167-8655},
year = {2020},
date = {2020-01-01},
journal = {Pattern Recognition Letters},
volume = {131},
pages = {361–367},
abstract = {Cutting-edge visualization and interaction technologies are increasingly used in museum exhibitions, providing novel ways to engage visitors and enhance their cultural experience. Existing applications are commonly built upon a single technology, focusing on visualization, motion or verbal interaction (e.g., high-resolution projections, gesture interfaces, chatbots). This aspect limits their potential, since museums are highly heterogeneous in terms of visitors profiles and interests, requiring multi-channel, customizable interaction modalities. To this aim, this work describes and evaluates an artificial intelligence powered, interactive holographic stand aimed at describing Leonardo Da Vinci’s art. This system provides the users with accurate 3D representations of Leonardo’s machines, which can be interactively manipulated through a touchless user interface. It is also able to dialog with the users in natural language about Leonardo’s art, while keeping the context of conversation and interactions. Furthermore, the results of a large user study, carried out during art and tech exhibitions, are presented and discussed. The goal was to assess how users of different ages and interests perceive, understand and explore cultural objects when holograms and artificial intelligence are used as instruments of knowledge and analysis.},
keywords = {Artificial Intelligence, Conversational systems, Cultural Heritage, Holograms, Human Computer Interaction, Touchless interaction, User study},
pubstate = {published},
tppubtype = {article}
}
Cutting-edge visualization and interaction technologies are increasingly used in museum exhibitions, providing novel ways to engage visitors and enhance their cultural experience. Existing applications are commonly built upon a single technology, focusing on visualization, motion or verbal interaction (e.g., high-resolution projections, gesture interfaces, chatbots). This aspect limits their potential, since museums are highly heterogeneous in terms of visitors profiles and interests, requiring multi-channel, customizable interaction modalities. To this aim, this work describes and evaluates an artificial intelligence powered, interactive holographic stand aimed at describing Leonardo Da Vinci’s art. This system provides the users with accurate 3D representations of Leonardo’s machines, which can be interactively manipulated through a touchless user interface. It is also able to dialog with the users in natural language about Leonardo’s art, while keeping the context of conversation and interactions. Furthermore, the results of a large user study, carried out during art and tech exhibitions, are presented and discussed. The goal was to assess how users of different ages and interests perceive, understand and explore cultural objects when holograms and artificial intelligence are used as instruments of knowledge and analysis.
Agostinelli, Sofia; Cumo, Fabrizio; Guidi, Giambattista; Tomazzoli, Claudio
The Potential of Digital Twin Model Integrated with Artificial Intelligence Systems Proceedings Article
In: 2020 IEEE International Conference on Environment and Electrical Engineering and 2020 IEEE Industrial and Commercial Power Systems Europe, EEEIC / I and CPS Europe 2020, pp. 1–6, Institute of Electrical and Electronics Engineers Inc., 2020, ISBN: 978-1-7281-7453-2.
Abstract | Links | BibTeX | Tags: Digital Twin, Edge computing, Energy Efficiency, Energy management, Machine Learning
@inproceedings{agostinelli_potential_2020,
title = {The Potential of Digital Twin Model Integrated with Artificial Intelligence Systems},
author = {Sofia Agostinelli and Fabrizio Cumo and Giambattista Guidi and Claudio Tomazzoli},
doi = {10.1109/EEEIC/ICPSEurope49358.2020.9160810},
isbn = {978-1-7281-7453-2},
year = {2020},
date = {2020-01-01},
booktitle = {2020 IEEE International Conference on Environment and Electrical Engineering and 2020 IEEE Industrial and Commercial Power Systems Europe, EEEIC / I and CPS Europe 2020},
pages = {1–6},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {The paper explores the use of a “digital twin model” applied to the case study of a residential district, and organized as a three-dimensional data system able to participate to the intelligent optimization and automation of the energy management and efficiency of the building system. The case study focuses on the area called Rinascimento III in Rome, consisting of 16 eight-floor building hosting 216 apartment units with an overall percentage of self-renewable energy produced by the building complex equal to 70%. This already quite high percentage means that the building complex can be defined as a Near Zero Energy Building (NZEB), i.e. a building that has a very high energy performance, and the nearly-zero or very low amount of energy required should be covered to a very significant extent by energy from renewable sources, including energy from renewable source produced on-site or nearby.},
keywords = {Digital Twin, Edge computing, Energy Efficiency, Energy management, Machine Learning},
pubstate = {published},
tppubtype = {inproceedings}
}
The paper explores the use of a “digital twin model” applied to the case study of a residential district, and organized as a three-dimensional data system able to participate to the intelligent optimization and automation of the energy management and efficiency of the building system. The case study focuses on the area called Rinascimento III in Rome, consisting of 16 eight-floor building hosting 216 apartment units with an overall percentage of self-renewable energy produced by the building complex equal to 70%. This already quite high percentage means that the building complex can be defined as a Near Zero Energy Building (NZEB), i.e. a building that has a very high energy performance, and the nearly-zero or very low amount of energy required should be covered to a very significant extent by energy from renewable sources, including energy from renewable source produced on-site or nearby.
Cristani, Matteo; Tomazzoli, Claudio
Dataset Anonymization on Cloud: Open Problems and Perspectives Proceedings Article
In: pp. 74–85, 2020.
Abstract | Links | BibTeX | Tags: Business Processes, Machine Learning, Natural Language Processing
@inproceedings{cristani_dataset_2020,
title = {Dataset Anonymization on Cloud: Open Problems and Perspectives},
author = {Matteo Cristani and Claudio Tomazzoli},
doi = {10.1007/978-3-030-51253-8_9},
year = {2020},
date = {2020-01-01},
volume = {11609 LNCS},
pages = {74–85},
series = {Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)},
abstract = {Data anonymization is the process of making information contained in a group of data such that it is not possible to identify unique references to single elements in the group after the process. This action, when conducted onto datasets used to make statistical inference is bound to have ananlogous behaviours on certain indices before and after the process itself. In this paper we study the pipeline of anonymization process for datasets, when this pipeline is managed on cloud technology, where cryptography is not applicable at all, for datasets being available in an open setting. We examine the open problems, and devise a method to address these problems in a logical framework.},
keywords = {Business Processes, Machine Learning, Natural Language Processing},
pubstate = {published},
tppubtype = {inproceedings}
}
Data anonymization is the process of making information contained in a group of data such that it is not possible to identify unique references to single elements in the group after the process. This action, when conducted onto datasets used to make statistical inference is bound to have ananlogous behaviours on certain indices before and after the process itself. In this paper we study the pipeline of anonymization process for datasets, when this pipeline is managed on cloud technology, where cryptography is not applicable at all, for datasets being available in an open setting. We examine the open problems, and devise a method to address these problems in a logical framework.
Cristani, Matteo; Pasetto, Luca; Tomazzoli, Claudio
A Knowledge-Intensive Methodology for Explainable Sales Prediction Proceedings Article
In: pp. 1180–1187, 2020.
Abstract | Links | BibTeX | Tags: Automatic reasoning, Business Processes, Machine Learning
@inproceedings{cristani_knowledge-intensive_2020,
title = {A Knowledge-Intensive Methodology for Explainable Sales Prediction},
author = {Matteo Cristani and Luca Pasetto and Claudio Tomazzoli},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85093363815&doi=10.1016%2fj.procs.2020.09.114&partnerID=40&md5=abd8bf5a398a7d9c62865678c6e6d97e},
doi = {10.1016/j.procs.2020.09.114},
year = {2020},
date = {2020-01-01},
volume = {176},
pages = {1180–1187},
series = {Procedia Computer Science},
abstract = {Sales prediction in food market is a complex issue that has been addressed in the recent past with machine learning techniques. Although some promising results, an experimental work that we describe in this paper shows some drawbacks of the above mentioned data-driven method and habilitates the definition of a novel methodology, strongly involving a piori knowledge.},
keywords = {Automatic reasoning, Business Processes, Machine Learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Sales prediction in food market is a complex issue that has been addressed in the recent past with machine learning techniques. Although some promising results, an experimental work that we describe in this paper shows some drawbacks of the above mentioned data-driven method and habilitates the definition of a novel methodology, strongly involving a piori knowledge.
2019
Pecorelli, Fabiano; Nucci, Dario Di; Roover, Coen De; Lucia, Andrea De
On the role of data balancing for machine learning-based code smell detection Proceedings Article
In: Proceedings of the 3rd ACM SIGSOFT International Workshop on Machine Learning Techniques for Software Quality Evaluation, pp. 19–24, ACM, Tallinn Estonia, 2019, ISBN: 978-1-4503-6855-1.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Code Smell Detection, Empirical Software Engineering, Technical Debt Management
@inproceedings{pecorelliRoleDataBalancing2019,
title = {On the role of data balancing for machine learning-based code smell detection},
author = {Fabiano Pecorelli and Dario Di Nucci and Coen De Roover and Andrea De Lucia},
url = {https://dl.acm.org/doi/10.1145/3340482.3342744},
doi = {10.1145/3340482.3342744},
isbn = {978-1-4503-6855-1},
year = {2019},
date = {2019-08-01},
urldate = {2024-07-07},
booktitle = {Proceedings of the 3rd ACM SIGSOFT International Workshop on Machine Learning Techniques for Software Quality Evaluation},
pages = {19–24},
publisher = {ACM},
address = {Tallinn Estonia},
abstract = {Code smells can compromise software quality in the long term by inducing technical debt. For this reason, many approaches aimed at identifying these design flaws have been proposed in the last decade. Most of them are based on heuristics in which a set of metrics (e.g., code metrics, process metrics) is used to detect smelly code components. However, these techniques suffer of subjective interpretation, low agreement between detectors, and threshold dependability. To overcome these limitations, previous work applied Machine Learning techniques that can learn from previous datasets without needing any threshold definition. However, more recent work has shown that Machine Learning is not always suitable for code smell detection due to the highly unbalanced nature of the problem. In this study we investigate several approaches able to mitigate data unbalancing issues to understand their impact on ML-based approaches for code smell detection. Our findings highlight a number of limitations and open issues with respect to the usage of data balancing in ML-based code smell detection.},
keywords = {Artificial Intelligence, Code Smell Detection, Empirical Software Engineering, Technical Debt Management},
pubstate = {published},
tppubtype = {inproceedings}
}
Code smells can compromise software quality in the long term by inducing technical debt. For this reason, many approaches aimed at identifying these design flaws have been proposed in the last decade. Most of them are based on heuristics in which a set of metrics (e.g., code metrics, process metrics) is used to detect smelly code components. However, these techniques suffer of subjective interpretation, low agreement between detectors, and threshold dependability. To overcome these limitations, previous work applied Machine Learning techniques that can learn from previous datasets without needing any threshold definition. However, more recent work has shown that Machine Learning is not always suitable for code smell detection due to the highly unbalanced nature of the problem. In this study we investigate several approaches able to mitigate data unbalancing issues to understand their impact on ML-based approaches for code smell detection. Our findings highlight a number of limitations and open issues with respect to the usage of data balancing in ML-based code smell detection.
Essmaeel, Kyis; Migniot, Cyrille; Dipanda, Albert; Gallo, Luigi; Damiani, Ernesto; Pietro, Giuseppe De
A new 3D descriptor for human classification: application for human detection in a multi-kinect system Journal Article
In: Multimedia Tools and Applications, vol. 78, no. 16, pp. 22479–22508, 2019, ISSN: 1573-7721.
Abstract | Links | BibTeX | Tags: 3D descriptor, Artificial Intelligence, Classification, Human detection, Kinect
@article{essmaeelNew3DDescriptor2019,
title = {A new 3D descriptor for human classification: application for human detection in a multi-kinect system},
author = {Kyis Essmaeel and Cyrille Migniot and Albert Dipanda and Luigi Gallo and Ernesto Damiani and Giuseppe De Pietro},
url = {https://doi.org/10.1007/s11042-019-7568-6},
doi = {10.1007/s11042-019-7568-6},
issn = {1573-7721},
year = {2019},
date = {2019-08-01},
journal = {Multimedia Tools and Applications},
volume = {78},
number = {16},
pages = {22479–22508},
abstract = {In this paper we present a new 3D descriptor for human classification and a human detection method based on this descriptor. The proposed 3D descriptor allows classification of an object represented by a point cloud, as human or non-human. It is derived from the well-known Histogram of Oriented Gradient by employing surface normals instead of gradients. The process consists in an appropriate subdivision of the object point cloud into blocks. These blocks provide the spatial distribution modeling of the surface normal orientation into the different parts of the object. This distribution modelling is expressed as a histogram. In addition we have set up a multi-kinect acquisition system that provides us with Complete Point Clouds (CPC) (i.e. 360° view). Such CPCs enable a suitable processing, particularly in case of occlusions. Moreover they allow for the determination of the human frontal orientation. Based on the proposed 3D descriptor, we have developed a human detection method that is applied on CPCs. First, we evaluated the 3D descriptor over a set of CPC candidates by using the Support Vector Machine (SVM) classifier. The learning process was conducted with the original CPC database that we have built. The results are very promising. The descriptor can discriminate human from non-human candidates and provides the frontal direction of humans with high precision. In addition we demonstrated that using the CPCs improves significantly the classification results in comparison with Single Point Clouds (i.e. points clouds acquired with only one kinect). Second, we compared our detection method with two others, namely the HOG detector on RGB images and a 3D HOG-based detection method that is applied on RGB-depth data. The obtained results on different situations show that the proposed human detection method provides excellent performances that outperform the other two detection methods.},
keywords = {3D descriptor, Artificial Intelligence, Classification, Human detection, Kinect},
pubstate = {published},
tppubtype = {article}
}
In this paper we present a new 3D descriptor for human classification and a human detection method based on this descriptor. The proposed 3D descriptor allows classification of an object represented by a point cloud, as human or non-human. It is derived from the well-known Histogram of Oriented Gradient by employing surface normals instead of gradients. The process consists in an appropriate subdivision of the object point cloud into blocks. These blocks provide the spatial distribution modeling of the surface normal orientation into the different parts of the object. This distribution modelling is expressed as a histogram. In addition we have set up a multi-kinect acquisition system that provides us with Complete Point Clouds (CPC) (i.e. 360° view). Such CPCs enable a suitable processing, particularly in case of occlusions. Moreover they allow for the determination of the human frontal orientation. Based on the proposed 3D descriptor, we have developed a human detection method that is applied on CPCs. First, we evaluated the 3D descriptor over a set of CPC candidates by using the Support Vector Machine (SVM) classifier. The learning process was conducted with the original CPC database that we have built. The results are very promising. The descriptor can discriminate human from non-human candidates and provides the frontal direction of humans with high precision. In addition we demonstrated that using the CPCs improves significantly the classification results in comparison with Single Point Clouds (i.e. points clouds acquired with only one kinect). Second, we compared our detection method with two others, namely the HOG detector on RGB images and a 3D HOG-based detection method that is applied on RGB-depth data. The obtained results on different situations show that the proposed human detection method provides excellent performances that outperform the other two detection methods.
Essmaeel, Kyis; Migniot, Cyrille; Dipanda, Albert; Gallo, Luigi; Damiani, Ernesto; Pietro, Giuseppe De
A new 3D descriptor for human classification: application for human detection in a multi-kinect system Journal Article
In: Multimedia Tools and Applications, vol. 78, no. 16, pp. 22479–22508, 2019, ISSN: 1573-7721.
Abstract | Links | BibTeX | Tags: 3D descriptor, Artificial Intelligence, Classification, Human detection, Kinect
@article{essmaeel_new_2019,
title = {A new 3D descriptor for human classification: application for human detection in a multi-kinect system},
author = {Kyis Essmaeel and Cyrille Migniot and Albert Dipanda and Luigi Gallo and Ernesto Damiani and Giuseppe De Pietro},
url = {https://doi.org/10.1007/s11042-019-7568-6},
doi = {10.1007/s11042-019-7568-6},
issn = {1573-7721},
year = {2019},
date = {2019-08-01},
journal = {Multimedia Tools and Applications},
volume = {78},
number = {16},
pages = {22479–22508},
abstract = {In this paper we present a new 3D descriptor for human classification and a human detection method based on this descriptor. The proposed 3D descriptor allows classification of an object represented by a point cloud, as human or non-human. It is derived from the well-known Histogram of Oriented Gradient by employing surface normals instead of gradients. The process consists in an appropriate subdivision of the object point cloud into blocks. These blocks provide the spatial distribution modeling of the surface normal orientation into the different parts of the object. This distribution modelling is expressed as a histogram. In addition we have set up a multi-kinect acquisition system that provides us with Complete Point Clouds (CPC) (i.e. 360° view). Such CPCs enable a suitable processing, particularly in case of occlusions. Moreover they allow for the determination of the human frontal orientation. Based on the proposed 3D descriptor, we have developed a human detection method that is applied on CPCs. First, we evaluated the 3D descriptor over a set of CPC candidates by using the Support Vector Machine (SVM) classifier. The learning process was conducted with the original CPC database that we have built. The results are very promising. The descriptor can discriminate human from non-human candidates and provides the frontal direction of humans with high precision. In addition we demonstrated that using the CPCs improves significantly the classification results in comparison with Single Point Clouds (i.e. points clouds acquired with only one kinect). Second, we compared our detection method with two others, namely the HOG detector on RGB images and a 3D HOG-based detection method that is applied on RGB-depth data. The obtained results on different situations show that the proposed human detection method provides excellent performances that outperform the other two detection methods.},
keywords = {3D descriptor, Artificial Intelligence, Classification, Human detection, Kinect},
pubstate = {published},
tppubtype = {article}
}
In this paper we present a new 3D descriptor for human classification and a human detection method based on this descriptor. The proposed 3D descriptor allows classification of an object represented by a point cloud, as human or non-human. It is derived from the well-known Histogram of Oriented Gradient by employing surface normals instead of gradients. The process consists in an appropriate subdivision of the object point cloud into blocks. These blocks provide the spatial distribution modeling of the surface normal orientation into the different parts of the object. This distribution modelling is expressed as a histogram. In addition we have set up a multi-kinect acquisition system that provides us with Complete Point Clouds (CPC) (i.e. 360° view). Such CPCs enable a suitable processing, particularly in case of occlusions. Moreover they allow for the determination of the human frontal orientation. Based on the proposed 3D descriptor, we have developed a human detection method that is applied on CPCs. First, we evaluated the 3D descriptor over a set of CPC candidates by using the Support Vector Machine (SVM) classifier. The learning process was conducted with the original CPC database that we have built. The results are very promising. The descriptor can discriminate human from non-human candidates and provides the frontal direction of humans with high precision. In addition we demonstrated that using the CPCs improves significantly the classification results in comparison with Single Point Clouds (i.e. points clouds acquired with only one kinect). Second, we compared our detection method with two others, namely the HOG detector on RGB images and a 3D HOG-based detection method that is applied on RGB-depth data. The obtained results on different situations show that the proposed human detection method provides excellent performances that outperform the other two detection methods.
Pecorelli, Fabiano; Palomba, Fabio; Nucci, Dario Di; Lucia, Andrea De
Comparing Heuristic and Machine Learning Approaches for Metric-Based Code Smell Detection Proceedings Article
In: 2019 IEEE/ACM 27th International Conference on Program Comprehension (ICPC), pp. 93–104, IEEE, Montreal, QC, Canada, 2019, ISBN: 978-1-7281-1519-1.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Code Smell Detection, Software Engineering, Technical Debt Management
@inproceedings{pecorelliComparingHeuristicMachine2019,
title = {Comparing Heuristic and Machine Learning Approaches for Metric-Based Code Smell Detection},
author = {Fabiano Pecorelli and Fabio Palomba and Dario Di Nucci and Andrea De Lucia},
url = {https://ieeexplore.ieee.org/document/8813271/},
doi = {10.1109/ICPC.2019.00023},
isbn = {978-1-7281-1519-1},
year = {2019},
date = {2019-05-01},
urldate = {2024-07-07},
booktitle = {2019 IEEE/ACM 27th International Conference on Program Comprehension (ICPC)},
pages = {93–104},
publisher = {IEEE},
address = {Montreal, QC, Canada},
abstract = {Code smells represent poor implementation choices performed by developers when enhancing source code. Their negative impact on source code maintainability and comprehensibility has been widely shown in the past and several techniques to automatically detect them have been devised. Most of these techniques are based on heuristics, namely they compute a set of code metrics and combine them by creating detection rules; while they have a reasonable accuracy, a recent trend is represented by the use of machine learning where code metrics are used as predictors of the smelliness of code artefacts. Despite the recent advances in the field, there is still a noticeable lack of knowledge of whether machine learning can actually be more accurate than traditional heuristic-based approaches. To fill this gap, in this paper we propose a large-scale study to empirically compare the performance of heuristic-based and machine-learning-based techniques for metric-based code smell detection. We consider five code smell types and compare machine learning models with DECOR, a state-of-the-art heuristic-based approach. Key findings emphasize the need of further research aimed at improving the effectiveness of both machine learning and heuristic approaches for code smell detection: while DECOR generally achieves better performance than a machine learning baseline, its precision is still too low to make it usable in practice.},
keywords = {Artificial Intelligence, Code Smell Detection, Software Engineering, Technical Debt Management},
pubstate = {published},
tppubtype = {inproceedings}
}
Code smells represent poor implementation choices performed by developers when enhancing source code. Their negative impact on source code maintainability and comprehensibility has been widely shown in the past and several techniques to automatically detect them have been devised. Most of these techniques are based on heuristics, namely they compute a set of code metrics and combine them by creating detection rules; while they have a reasonable accuracy, a recent trend is represented by the use of machine learning where code metrics are used as predictors of the smelliness of code artefacts. Despite the recent advances in the field, there is still a noticeable lack of knowledge of whether machine learning can actually be more accurate than traditional heuristic-based approaches. To fill this gap, in this paper we propose a large-scale study to empirically compare the performance of heuristic-based and machine-learning-based techniques for metric-based code smell detection. We consider five code smell types and compare machine learning models with DECOR, a state-of-the-art heuristic-based approach. Key findings emphasize the need of further research aimed at improving the effectiveness of both machine learning and heuristic approaches for code smell detection: while DECOR generally achieves better performance than a machine learning baseline, its precision is still too low to make it usable in practice.
Mengoni, Maura; Generosi, Andrea; Giraldi, Luca; Torcianti, Marco
A method to measure the emotional experience of audience by the EMOJ tool. The case study of Macerata Opera Festival Journal Article
In: Marketing i Rynek, no. 10, pp. 4–13, 2019, ISSN: 1231-7853.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Emotion Recognition, Tourism, User experience
@article{mengoni_method_2019,
title = {A method to measure the emotional experience of audience by the EMOJ tool. The case study of Macerata Opera Festival},
author = {Maura Mengoni and Andrea Generosi and Luca Giraldi and Marco Torcianti},
url = {http://cejsh.icm.edu.pl/cejsh/element/bwmeta1.element.ojs-doi-10_33226_1231-7853_2019_10_1},
doi = {10.33226/1231-7853.2019.10.1},
issn = {1231-7853},
year = {2019},
date = {2019-01-01},
urldate = {2024-12-28},
journal = {Marketing i Rynek},
number = {10},
pages = {4–13},
abstract = {This paper aims to present a case study on the application of Emotional Analytics to measure audience experience in the culture sector. The adopted technology enables audience measurement by detecting persons' face and recognizing the emotions they feel in real time, while watching a show or attending a cultural event. It is the result of a long-term research and development project, whose goal is to advance neuro-marketing by proving a non-invasive ad wearable technology to investigate individual affective and emotional response in public spaces. The developed Emotional Analytics platform is called EMOJ and in summer 2019 has been used to analyse the experience lived by the audience of the Macerata Opera Festival, a series of opera representations that take place in the Sferisterio Arena, in Macerata. The goal of this project is to provide useful information on the quality of each performance and of the entire festival perceived by the audience, in order to make the right choices to improve the performances and to have a return on ticket sales for the coming years.},
keywords = {Artificial Intelligence, Emotion Recognition, Tourism, User experience},
pubstate = {published},
tppubtype = {article}
}
This paper aims to present a case study on the application of Emotional Analytics to measure audience experience in the culture sector. The adopted technology enables audience measurement by detecting persons' face and recognizing the emotions they feel in real time, while watching a show or attending a cultural event. It is the result of a long-term research and development project, whose goal is to advance neuro-marketing by proving a non-invasive ad wearable technology to investigate individual affective and emotional response in public spaces. The developed Emotional Analytics platform is called EMOJ and in summer 2019 has been used to analyse the experience lived by the audience of the Macerata Opera Festival, a series of opera representations that take place in the Sferisterio Arena, in Macerata. The goal of this project is to provide useful information on the quality of each performance and of the entire festival perceived by the audience, in order to make the right choices to improve the performances and to have a return on ticket sales for the coming years.
Galteri, Leonardo; Seidenari, Lorenzo; Bertini, Marco; Bimbo, Alberto Del
Deep universal generative adversarial compression artifact removal Journal Article
In: IEEE Transactions on Multimedia, vol. 21, no. 8, pp. 2131–2145, 2019, (ISBN: 1520-9210 tex.copyright: All rights reserved).
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Compression Artifact Removal, Computer Vision and Pattern Recognition, Generative Adversarial Networks, Image Compression, Image Processing, Image Synthesis and Enhancement
@article{galteriDeepUniversalGenerative2019,
title = {Deep universal generative adversarial compression artifact removal},
author = {Leonardo Galteri and Lorenzo Seidenari and Marco Bertini and Alberto Del Bimbo},
url = {https://ieeexplore.ieee.org/abstract/document/8625533/},
doi = {10.1109/TMM.2019.2895280},
year = {2019},
date = {2019-01-01},
journal = {IEEE Transactions on Multimedia},
volume = {21},
number = {8},
pages = {2131–2145},
publisher = {IEEE},
abstract = {Image compression is a need that arises in many circumstances. Unfortunately, whenever a lossy compression algorithm is used, artifacts will manifest. Image artifacts, caused by compression tend to eliminate higher frequency details and, in certain cases, may add noise or small image structures. There are two main drawbacks of this phenomenon. First, images appear much less pleasant to the human eye. Second, computer vision algorithms, such as object detectors, may be hindered and their performance reduced. Removing such artifacts means recovering the original image from a perturbed version of it. This means that one ideally should invert the compression process through a complicated nonlinear image transformation. We propose an image transformation approach based on a feedforward fully convolutional residual network model. We show that this model can be optimized either traditionally, directly optimizing an image similarity loss (SSIM), or using a generative adversarial approach (GAN). Our GAN is able to produce images with more photorealistic details than SSIM-based networks. We describe a novel training procedure based on subpatches and devise a novel testing protocol to evaluate restored images quantitatively. We show that our approach can be used as a preprocessing step for different computer vision tasks in case images are degraded by compression to a point that state-of-the art algorithms fail. In this case, our GAN-based approach obtains better performance than MSE or SSIM trained networks. Different from previously proposed approaches, we are able to remove artifacts generated at any QF by inferring the image quality directly from data.},
note = {ISBN: 1520-9210
tex.copyright: All rights reserved},
keywords = {Artificial Intelligence, Compression Artifact Removal, Computer Vision and Pattern Recognition, Generative Adversarial Networks, Image Compression, Image Processing, Image Synthesis and Enhancement},
pubstate = {published},
tppubtype = {article}
}
Image compression is a need that arises in many circumstances. Unfortunately, whenever a lossy compression algorithm is used, artifacts will manifest. Image artifacts, caused by compression tend to eliminate higher frequency details and, in certain cases, may add noise or small image structures. There are two main drawbacks of this phenomenon. First, images appear much less pleasant to the human eye. Second, computer vision algorithms, such as object detectors, may be hindered and their performance reduced. Removing such artifacts means recovering the original image from a perturbed version of it. This means that one ideally should invert the compression process through a complicated nonlinear image transformation. We propose an image transformation approach based on a feedforward fully convolutional residual network model. We show that this model can be optimized either traditionally, directly optimizing an image similarity loss (SSIM), or using a generative adversarial approach (GAN). Our GAN is able to produce images with more photorealistic details than SSIM-based networks. We describe a novel training procedure based on subpatches and devise a novel testing protocol to evaluate restored images quantitatively. We show that our approach can be used as a preprocessing step for different computer vision tasks in case images are degraded by compression to a point that state-of-the art algorithms fail. In this case, our GAN-based approach obtains better performance than MSE or SSIM trained networks. Different from previously proposed approaches, we are able to remove artifacts generated at any QF by inferring the image quality directly from data.
Berlincioni, Lorenzo; Becattini, Federico; Galteri, Leonardo; Seidenari, Lorenzo; Bimbo, Alberto Del
Road layout understanding by generative adversarial inpainting Proceedings Article
In: Inpainting and Denoising Challenges, pp. 111–128, Springer International Publishing, 2019, ISBN: 3-030-25613-8, (tex.copyright: All rights reserved).
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Autonomous Driving, Computer Vision and Pattern Recognition, Generative Adversarial Networks, Image Inpainting, Image Processing, Semantic Segmentation, Urban Planning
@inproceedings{berlincioniRoadLayoutUnderstanding2019,
title = {Road layout understanding by generative adversarial inpainting},
author = {Lorenzo Berlincioni and Federico Becattini and Leonardo Galteri and Lorenzo Seidenari and Alberto Del Bimbo},
url = {https://link.springer.com/chapter/10.1007/978-3-030-25614-2_10},
doi = {10.1007/978-3-030-25614-2_10},
isbn = {3-030-25613-8},
year = {2019},
date = {2019-01-01},
booktitle = {Inpainting and Denoising Challenges},
pages = {111–128},
publisher = {Springer International Publishing},
abstract = {Autonomous driving is becoming a reality, yet vehicles still need to rely on complex sensor fusion to understand the scene they act in. The ability to discern static environment and dynamic entities provides a comprehension of the road layout that poses constraints to the reasoning process about moving objects. We pursue this through a GAN-based semantic segmentation inpainting model to remove all dynamic objects from the scene and focus on understanding its static components such as streets, sidewalks and buildings. We evaluate this task on the Cityscapes dataset and on a novel synthetically generated dataset obtained with the CARLA simulator and specifically designed to quantitatively evaluate semantic segmentation inpaintings. We compare our methods with a variety of baselines working both in the RGB and segmentation domains.},
note = {tex.copyright: All rights reserved},
keywords = {Artificial Intelligence, Autonomous Driving, Computer Vision and Pattern Recognition, Generative Adversarial Networks, Image Inpainting, Image Processing, Semantic Segmentation, Urban Planning},
pubstate = {published},
tppubtype = {inproceedings}
}
Autonomous driving is becoming a reality, yet vehicles still need to rely on complex sensor fusion to understand the scene they act in. The ability to discern static environment and dynamic entities provides a comprehension of the road layout that poses constraints to the reasoning process about moving objects. We pursue this through a GAN-based semantic segmentation inpainting model to remove all dynamic objects from the scene and focus on understanding its static components such as streets, sidewalks and buildings. We evaluate this task on the Cityscapes dataset and on a novel synthetically generated dataset obtained with the CARLA simulator and specifically designed to quantitatively evaluate semantic segmentation inpaintings. We compare our methods with a variety of baselines working both in the RGB and segmentation domains.
Turchini, Francesco; Seidenari, Lorenzo; Galteri, Leonardo; Ferracani, Andrea; Becchi, Giuseppe; Bimbo, Alberto Del
Flexible automatic football filming and summarization Proceedings Article
In: Proceedings Proceedings of the 2nd International Workshop on Multimedia Content Analysis in Sports, pp. 108–114, 2019, (tex.copyright: All rights reserved).
Abstract | Links | BibTeX | Tags: Activity Recognition, Artificial Intelligence, Automatic Highlight Detection, Computer Vision and Pattern Recognition, Information Systems, Real-time Video Processing, Sports Broadcasting, Sports Video Analysis, Video Summarization
@inproceedings{turchiniFlexibleAutomaticFootball2019,
title = {Flexible automatic football filming and summarization},
author = {Francesco Turchini and Lorenzo Seidenari and Leonardo Galteri and Andrea Ferracani and Giuseppe Becchi and Alberto Del Bimbo},
url = {https://dl.acm.org/doi/10.1145/3347318.3355526},
doi = {10.1145/3347318.3355526},
year = {2019},
date = {2019-01-01},
booktitle = {Proceedings Proceedings of the 2nd International Workshop on Multimedia Content Analysis in Sports},
pages = {108–114},
abstract = {We propose a method aimed at reducing human intervention in football video shooting and highlights editing, allowing automatic highlight detection together with panning and zooming on salient areas of the playing field. Our recognition subsystem exploits computer vision algorithms to perform automatic detection, pan and zoom and extraction of salient segments of a recorded match. Matches are elaborated offline, extracting and analyzing motion and visual features of the elements in salient zones of the scene, i.e. midfield circle and penalty areas. Automatic summarization is performed by classifying subsequences of a match with machine learning algorithms, which are pretrained on previously acquired and annotated videos of other matches. Among salient actions, special attention is given to goal events, but also other generic highlights are identified. The only assumption for our method to work is to employ a pair of cameras which should frame the football pitch splitting the field in two halves. We demonstrate the functioning of our approach using two ultra high definition cameras, building a system which is also able to collect various metadata of the matches to extrapolate other salient information.},
note = {tex.copyright: All rights reserved},
keywords = {Activity Recognition, Artificial Intelligence, Automatic Highlight Detection, Computer Vision and Pattern Recognition, Information Systems, Real-time Video Processing, Sports Broadcasting, Sports Video Analysis, Video Summarization},
pubstate = {published},
tppubtype = {inproceedings}
}
We propose a method aimed at reducing human intervention in football video shooting and highlights editing, allowing automatic highlight detection together with panning and zooming on salient areas of the playing field. Our recognition subsystem exploits computer vision algorithms to perform automatic detection, pan and zoom and extraction of salient segments of a recorded match. Matches are elaborated offline, extracting and analyzing motion and visual features of the elements in salient zones of the scene, i.e. midfield circle and penalty areas. Automatic summarization is performed by classifying subsequences of a match with machine learning algorithms, which are pretrained on previously acquired and annotated videos of other matches. Among salient actions, special attention is given to goal events, but also other generic highlights are identified. The only assumption for our method to work is to employ a pair of cameras which should frame the football pitch splitting the field in two halves. We demonstrate the functioning of our approach using two ultra high definition cameras, building a system which is also able to collect various metadata of the matches to extrapolate other salient information.
Becattini, Federico; Seidenari, Lorenzo; Berlincioni, Lorenzo; Galteri, Leonardo; Bimbo, Alberto Del
Vehicle trajectories from unlabeled data through iterative plane registration Proceedings Article
In: Image Analysis and Processing–ICIAP 2019: 20th International Conference, Trento, Italy, September 9–13, 2019, Proceedings, Part I 20, pp. 60–70, Springer International Publishing, 2019, ISBN: 3-030-30641-0, (tex.copyright: All rights reserved).
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Autonomous Driving, Computer Vision and Pattern Recognition, Plane Registration, Semantic Segmentation, Trajectory Prediction, Urban Planning
@inproceedings{becattiniVehicleTrajectoriesUnlabeled2019,
title = {Vehicle trajectories from unlabeled data through iterative plane registration},
author = {Federico Becattini and Lorenzo Seidenari and Lorenzo Berlincioni and Leonardo Galteri and Alberto Del Bimbo},
url = {https://link.springer.com/chapter/10.1007/978-3-030-30642-7_6},
doi = {10.1007/978-3-030-30642-7_6},
isbn = {3-030-30641-0},
year = {2019},
date = {2019-01-01},
booktitle = {Image Analysis and Processing–ICIAP 2019: 20th International Conference, Trento, Italy, September 9–13, 2019, Proceedings, Part I 20},
pages = {60–70},
publisher = {Springer International Publishing},
abstract = {One of the most complex aspects of autonomous driving concerns understanding the surrounding environment. In particular, the interest falls on detecting which agents are populating it and how they are moving. The capacity to predict how these may act in the near future would allow an autonomous vehicle to safely plan its trajectory, minimizing the risks for itself and others. In this work we propose an automatic trajectory annotation method exploiting an Iterative Plane Registration algorithm based on homographies and semantic segmentations. The output of our technique is a set of holistic trajectories (past-present-future) paired with a single image context, useful to train a predictive model.},
note = {tex.copyright: All rights reserved},
keywords = {Artificial Intelligence, Autonomous Driving, Computer Vision and Pattern Recognition, Plane Registration, Semantic Segmentation, Trajectory Prediction, Urban Planning},
pubstate = {published},
tppubtype = {inproceedings}
}
One of the most complex aspects of autonomous driving concerns understanding the surrounding environment. In particular, the interest falls on detecting which agents are populating it and how they are moving. The capacity to predict how these may act in the near future would allow an autonomous vehicle to safely plan its trajectory, minimizing the risks for itself and others. In this work we propose an automatic trajectory annotation method exploiting an Iterative Plane Registration algorithm based on homographies and semantic segmentations. The output of our technique is a set of holistic trajectories (past-present-future) paired with a single image context, useful to train a predictive model.
Galteri, Leonardo; Seidenari, Lorenzo; Bertini, Marco; Bimbo, Alberto Del
Towards real-time image enhancement GANs Proceedings Article
In: Computer Analysis of Images and Patterns: 18th International Conference, CAIP 2019, Salerno, Italy, September 3–5, 2019, Proceedings, Part I 18, pp. 183–195, Springer International Publishing, 2019, ISBN: 3-030-29887-6, (tex.copyright: All rights reserved).
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Compression Artifact Removal, Computer Vision and Pattern Recognition, Generative Adversarial Networks, Image Synthesis and Enhancement, Real-time Video Processing, Video Summarization
@inproceedings{galteriRealtimeImageEnhancement2019,
title = {Towards real-time image enhancement GANs},
author = {Leonardo Galteri and Lorenzo Seidenari and Marco Bertini and Alberto Del Bimbo},
url = {https://link.springer.com/chapter/10.1007/978-3-030-29888-3_15},
doi = {10.1007/978-3-030-29888-3_15},
isbn = {3-030-29887-6},
year = {2019},
date = {2019-01-01},
booktitle = {Computer Analysis of Images and Patterns: 18th International Conference, CAIP 2019, Salerno, Italy, September 3–5, 2019, Proceedings, Part I 18},
pages = {183–195},
publisher = {Springer International Publishing},
abstract = {Video stream compression, using lossy algorithms, is performed to reduce the bandwidth required for transmission. To improve the video quality, either for human view or for automatic video analysis, videos are post-processed to eliminate the introduced compression artifacts. Generative Adversarial Network have been shown to obtain extremely high quality results in image enhancement tasks; however, to obtain top quality results high capacity large generators are usually employed, resulting in high computational costs and processing time. In this paper we present an architecture that can be used to reduce the cost of generators, paving a way towards real-time frame enhancement with GANs. With the proposed approach, enhanced images appear natural and pleasant to the eye. Locally high frequency patterns often differ from the raw uncompressed images. A possible application is to improve video conferencing, or live streaming. In these cases there is no original uncompressed video stream available. Therefore, we report results using popular no-reference metrics showing high naturalness and quality even for efficient networks.},
note = {tex.copyright: All rights reserved},
keywords = {Artificial Intelligence, Compression Artifact Removal, Computer Vision and Pattern Recognition, Generative Adversarial Networks, Image Synthesis and Enhancement, Real-time Video Processing, Video Summarization},
pubstate = {published},
tppubtype = {inproceedings}
}
Video stream compression, using lossy algorithms, is performed to reduce the bandwidth required for transmission. To improve the video quality, either for human view or for automatic video analysis, videos are post-processed to eliminate the introduced compression artifacts. Generative Adversarial Network have been shown to obtain extremely high quality results in image enhancement tasks; however, to obtain top quality results high capacity large generators are usually employed, resulting in high computational costs and processing time. In this paper we present an architecture that can be used to reduce the cost of generators, paving a way towards real-time frame enhancement with GANs. With the proposed approach, enhanced images appear natural and pleasant to the eye. Locally high frequency patterns often differ from the raw uncompressed images. A possible application is to improve video conferencing, or live streaming. In these cases there is no original uncompressed video stream available. Therefore, we report results using popular no-reference metrics showing high naturalness and quality even for efficient networks.
Galteri, Leonardo; Seidenari, Lorenzo; Bertini, Marco; Uricchio, Tiberio; Bimbo, Alberto Del
Fast video quality enhancement using gans Proceedings Article
In: Proceedings of the 27th ACM international conference on multimedia, pp. 1065–1067, 2019, (tex.copyright: All rights reserved).
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Compression Artifact Removal, Computer Vision and Pattern Recognition, Generative Adversarial Networks, Image Synthesis and Enhancement, Real-time Video Processing, Video Streaming
@inproceedings{galteriFastVideoQuality2019,
title = {Fast video quality enhancement using gans},
author = {Leonardo Galteri and Lorenzo Seidenari and Marco Bertini and Tiberio Uricchio and Alberto Del Bimbo},
url = {https://dl.acm.org/doi/abs/10.1145/3343031.3350592},
doi = {10.1145/3343031.3350592},
year = {2019},
date = {2019-01-01},
booktitle = {Proceedings of the 27th ACM international conference on multimedia},
pages = {1065–1067},
abstract = {Video compression algorithms result in a reduction of image quality, because of their lossy approach to reduce the required bandwidth. This affects commercial streaming services such as Netflix, or Amazon Prime Video, but affects also video conferencing and video surveillance systems. In all these cases it is possible to improve the video quality, both for human view and for automatic video analysis, without changing the compression pipeline, through a post-processing that eliminates the visual artifacts created by the compression algorithms. Generative Adversarial Networks have obtained extremely high quality results in image enhancement tasks; however, to obtain such results large generators are usually employed, resulting in high computational costs and processing time. In this work we present an architecture that can be used to reduce the computational cost and that has been implemented on mobile devices. A possible application is to improve video conferencing, or live streaming. In these cases there is no original uncompressed video stream available. Therefore, we report results using no-reference video quality metric showing high naturalness and quality even for efficient networks.},
note = {tex.copyright: All rights reserved},
keywords = {Artificial Intelligence, Compression Artifact Removal, Computer Vision and Pattern Recognition, Generative Adversarial Networks, Image Synthesis and Enhancement, Real-time Video Processing, Video Streaming},
pubstate = {published},
tppubtype = {inproceedings}
}
Video compression algorithms result in a reduction of image quality, because of their lossy approach to reduce the required bandwidth. This affects commercial streaming services such as Netflix, or Amazon Prime Video, but affects also video conferencing and video surveillance systems. In all these cases it is possible to improve the video quality, both for human view and for automatic video analysis, without changing the compression pipeline, through a post-processing that eliminates the visual artifacts created by the compression algorithms. Generative Adversarial Networks have obtained extremely high quality results in image enhancement tasks; however, to obtain such results large generators are usually employed, resulting in high computational costs and processing time. In this work we present an architecture that can be used to reduce the computational cost and that has been implemented on mobile devices. A possible application is to improve video conferencing, or live streaming. In these cases there is no original uncompressed video stream available. Therefore, we report results using no-reference video quality metric showing high naturalness and quality even for efficient networks.
Galteri, Leonardo; Ferrari, Claudio; Lisanti, Giuseppe; Berretti, Stefano; Bimbo, Alberto Del
Deep 3d morphable model refinement via progressive growing of conditional generative adversarial networks Journal Article
In: Computer Vision and Image Understanding, vol. 185, pp. 31–42, 2019, (ISBN: 1077-3142 tex.copyright: All rights reserved).
Abstract | Links | BibTeX | Tags: 3D Face Reconstruction, 3D Morphable Models, Artificial Intelligence, Computer Vision and Pattern Recognition, Digital Reconstruction, Generative Adversarial Networks, Image Synthesis and Enhancement
@article{galteriDeep3dMorphable2019,
title = {Deep 3d morphable model refinement via progressive growing of conditional generative adversarial networks},
author = {Leonardo Galteri and Claudio Ferrari and Giuseppe Lisanti and Stefano Berretti and Alberto Del Bimbo},
url = {https://www.sciencedirect.com/science/article/abs/pii/S1077314219300773},
doi = {10.1016/j.cviu.2019.05.002},
year = {2019},
date = {2019-01-01},
journal = {Computer Vision and Image Understanding},
volume = {185},
pages = {31–42},
publisher = {Academic Press},
abstract = {3D face reconstruction from a single 2D image is a fundamental Computer Vision problem of extraordinary difficulty. Statistical modeling techniques, such as the 3D Morphable Model (3DMM), have been widely exploited because of their capability of reconstructing a plausible model grounding on the prior knowledge of the facial shape. However, most of these techniques derive an approximated and smooth reconstruction of the face, without accounting for fine-grained details. In this work, we propose an approach based on a Conditional Generative Adversarial Network (CGAN) for refining the coarse reconstruction provided by a 3DMM. The latter is represented as a three channels image, where the pixel intensities represent the depth, curvature and elevation values of the 3D vertices. The architecture is an encoder–decoder, which is trained progressively, starting from the lower-resolution layers; this technique allows a more stable training, which leads to the generation of high quality outputs even when high-resolution images are fed during the training. Experimental results show that our method is able to produce reconstructions with fine-grained realistic details and lower reconstruction errors with respect to the 3DMM. A cross-dataset evaluation also shows that the network retains good generalization capabilities. Finally, comparison with state-of-the-art solutions evidence competitive performance, with comparable or lower error in most of the cases, and a clear improvement in the quality of the generated models.},
note = {ISBN: 1077-3142
tex.copyright: All rights reserved},
keywords = {3D Face Reconstruction, 3D Morphable Models, Artificial Intelligence, Computer Vision and Pattern Recognition, Digital Reconstruction, Generative Adversarial Networks, Image Synthesis and Enhancement},
pubstate = {published},
tppubtype = {article}
}
3D face reconstruction from a single 2D image is a fundamental Computer Vision problem of extraordinary difficulty. Statistical modeling techniques, such as the 3D Morphable Model (3DMM), have been widely exploited because of their capability of reconstructing a plausible model grounding on the prior knowledge of the facial shape. However, most of these techniques derive an approximated and smooth reconstruction of the face, without accounting for fine-grained details. In this work, we propose an approach based on a Conditional Generative Adversarial Network (CGAN) for refining the coarse reconstruction provided by a 3DMM. The latter is represented as a three channels image, where the pixel intensities represent the depth, curvature and elevation values of the 3D vertices. The architecture is an encoder–decoder, which is trained progressively, starting from the lower-resolution layers; this technique allows a more stable training, which leads to the generation of high quality outputs even when high-resolution images are fed during the training. Experimental results show that our method is able to produce reconstructions with fine-grained realistic details and lower reconstruction errors with respect to the 3DMM. A cross-dataset evaluation also shows that the network retains good generalization capabilities. Finally, comparison with state-of-the-art solutions evidence competitive performance, with comparable or lower error in most of the cases, and a clear improvement in the quality of the generated models.
Galteri, Leonardo; Ferrari, Claudio; Lisanti, Giuseppe; Berretti, Stefano; Bimbo, Alberto Del
Coarse-to-fine 3D face reconstruction Proceedings Article
In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, pp. 25–31, 2019, (tex.copyright: All rights reserved).
Abstract | Links | BibTeX | Tags: 3D Face Reconstruction, 3D Morphable Models, Artificial Intelligence, Computer Vision and Pattern Recognition, Digital Reconstruction, Generative Adversarial Networks, Image Synthesis and Enhancement
@inproceedings{galteriCoarsetofine3DFace2019,
title = {Coarse-to-fine 3D face reconstruction},
author = {Leonardo Galteri and Claudio Ferrari and Giuseppe Lisanti and Stefano Berretti and Alberto Del Bimbo},
url = {http://openaccess.thecvf.com/content_CVPRW_2019/papers/3D-WidDGET/Leonardo_Galteri_Coarse-to-Fine_3D_Face_Reconstruction_CVPRW_2019_paper.pdf},
doi = {Unavailable},
year = {2019},
date = {2019-01-01},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},
pages = {25–31},
abstract = {Reconstructing accurate 3D shapes of human faces from a single 2D image is a highly challenging Computer Vision problem that is studied since decades. Statistical modeling techniques, such as the 3D Morphable Model (3DMM), have been widely employed because of their capability of reconstructing a plausible model grounding on the prior knowledge of the facial shape. However, most of them derive a and smooth approximation of the real shape, without accounting for the surface details. In this work, we propose an approach based on a Conditional Generative Adversarial Network (CGAN) for refining the reconstruction provided by a 3DMM. The latter is represented as a three-channel image, where the pixel intensities represent, respectively, the depth and the azimuth and elevation angles of the surface normals. The network architecture is an encoder-decoder, which is trained progressively, starting from the lower-resolution layers; this technique allows a more stable training, which led to the generation of high-quality outputs even when high-resolution images are fed during the training. Experimental results show that our method is able to produce detailed realistic reconstructions and obtain lower errors with respect to the 3DMM. Finally, a comparison with a state-of-the-art solution evidences competitive performance and a clear improvement in the quality of the generated models.},
note = {tex.copyright: All rights reserved},
keywords = {3D Face Reconstruction, 3D Morphable Models, Artificial Intelligence, Computer Vision and Pattern Recognition, Digital Reconstruction, Generative Adversarial Networks, Image Synthesis and Enhancement},
pubstate = {published},
tppubtype = {inproceedings}
}
Reconstructing accurate 3D shapes of human faces from a single 2D image is a highly challenging Computer Vision problem that is studied since decades. Statistical modeling techniques, such as the 3D Morphable Model (3DMM), have been widely employed because of their capability of reconstructing a plausible model grounding on the prior knowledge of the facial shape. However, most of them derive a and smooth approximation of the real shape, without accounting for the surface details. In this work, we propose an approach based on a Conditional Generative Adversarial Network (CGAN) for refining the reconstruction provided by a 3DMM. The latter is represented as a three-channel image, where the pixel intensities represent, respectively, the depth and the azimuth and elevation angles of the surface normals. The network architecture is an encoder-decoder, which is trained progressively, starting from the lower-resolution layers; this technique allows a more stable training, which led to the generation of high-quality outputs even when high-resolution images are fed during the training. Experimental results show that our method is able to produce detailed realistic reconstructions and obtain lower errors with respect to the 3DMM. Finally, a comparison with a state-of-the-art solution evidences competitive performance and a clear improvement in the quality of the generated models.
Galteri, Leonardo; Ferrari, Claudio; Lisanti, Giuseppe; Berretti, Stefano; Bimbo, Alberto Del
Coarse to Fine 3D Face Reconstruction from Single Image Proceedings Article
In: 2019 14th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2019), pp. 1–1, IEEE, 2019, ISBN: 1-7281-0089-5, (tex.copyright: All rights reserved).
Abstract | Links | BibTeX | Tags: 3D Face Reconstruction, 3D Morphable Models, Artificial Intelligence, Computer Vision and Pattern Recognition, Digital Reconstruction, Generative Adversarial Networks, Image Synthesis and Enhancement
@inproceedings{galteriCoarseFine3D2019,
title = {Coarse to Fine 3D Face Reconstruction from Single Image},
author = {Leonardo Galteri and Claudio Ferrari and Giuseppe Lisanti and Stefano Berretti and Alberto Del Bimbo},
url = {https://ieeexplore.ieee.org/abstract/document/8756603},
doi = {10.1109/FG.2019.8756603},
isbn = {1-7281-0089-5},
year = {2019},
date = {2019-01-01},
booktitle = {2019 14th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2019)},
pages = {1–1},
publisher = {IEEE},
abstract = {In this demo we propose a coarse to fine reconstruction pipeline, which takes a single RGB image as input and outputs a detailed 3D model of the face. The pipeline is composed by two main blocks, the coarse reconstruction block, which is based on a 3D Morphable Model, and the refinement block, which instead grounds on a Generative Adversarial Network (GAN).},
note = {tex.copyright: All rights reserved},
keywords = {3D Face Reconstruction, 3D Morphable Models, Artificial Intelligence, Computer Vision and Pattern Recognition, Digital Reconstruction, Generative Adversarial Networks, Image Synthesis and Enhancement},
pubstate = {published},
tppubtype = {inproceedings}
}
In this demo we propose a coarse to fine reconstruction pipeline, which takes a single RGB image as input and outputs a detailed 3D model of the face. The pipeline is composed by two main blocks, the coarse reconstruction block, which is based on a 3D Morphable Model, and the refinement block, which instead grounds on a Generative Adversarial Network (GAN).