Organized by: Gianmarco Aversano, Sabri Shkiri, Euranova; and Yixi Xu, Microsoft
Machine Learning models are vulnerable to privacy attacks, and computer vision (CV) models are no exception to this. This means that simply publishing a pretrained image classifier or generator opens the door to attacks such as Membership Inference (MI), which aims to predict if a target image was part of the training set. Understanding why privacy attacks are possible and how to prevent them is crucial towards the objective of building Responsible AI tools.
In this workshop, we will cover the subject of privacy attacks in CV by introducing:
- Privacy in Machine Learning: what is it?
- The main Privacy Attacks against Machine Learning models
- How to defend Machine Learning models from Privacy Attacks
- Privacy vs Utility: optimizing the privacy-utility trade-off
- Privacy Auditing: how to estimate any privacy risk in advance?
- Invited speakers
- Hands-on sessions
We will also show practical examples using open-source libraries. Finally, we will also put extra focus on the synthetic image generators, which have been gaining momentum recently, and which we fear may be used wrongly.
We believe this workshop can be relevant for the following reasons:
- make the CV community aware of the privacy-utility trade-off that also exists in the image domain;
- provide CV practitioners with the theoretical and practical tools that are needed to estimate the privacy-utility trade-off in their use-case;
- given the increasing popularity of image generative models, this workshop can provide model-agnostic fundamentals on how to assess the privacy leakage of such models and the images they generate in a robust manner.
This workshop will be composed of keynote sessions on the theoretical background, keynote sessions for invited speakers whose papers have been accepted by this workshop, and one session on practical use-cases. The workshop will be held by Euranova's R&D department.
Euranova is a Belgian company who actively contributes to scientific research via their R&D department, which counts several publications in many AI-related domains. Euranova's research centre provides a clear view of the issues to address in the fields of artificial intelligence and data science.
Gianmarco Aversano is a Data Scientist in Euranova's Research Department. During his doctorate, he worked at CentraleSupélec Paris, University of Utah, and Université Libre de Bruxelles (ULB). His professional experience focuses on machine learning with a background in data engineering and MLOps, with both the knowledge gained from research and the knowledge acquired in the industrial field.
Yixi Xu is a senior applied research scientist at Microsoft AI for Good Lab mainly focusing on AI for health including its privacy risk estimation.
Sabri Shkiri leads the Euranova R&D department and managse internal research projects, back office R&D requests, technological watch for customers, technical assessments and innovation forum within Euranova.
Organized by: Christophoros NIKOU, University of Ioannina; Lazaros GRAMMATIKOPOULOS, Elli PETSA, Giorgos SFIKAS, University of West Attica; George RETSINAS, National Technical University of Athens; Panagiotis DIMITRAKOPOULOS, University of Ioannina; and Andreas EL SAER, University of West Attica
Photogrammetry and Computer Vision are two major fields with a significant overlap with Image Processing. We plan to invite presentations of high-quality works on topics that will include novel developments over classic Photogrammetric Computer Vision problems such as Structure from motion and SLAM, as well as papers with a focus on novel learning techniques over 3D geometric data. Topics of interest will include feature extraction, description and matching, multispectral and hyperspectral image processing and fusion, Multi-View Reconstruction and Surface Reconstruction, 3D point cloud analysis and processing, scene understanding, robot vision and perception, path and motion planning.
We believe that a workshop with a focus on Photogrammetry and Computer Vision will provide a very positive impact on the scope of ICIP, as well as attract more researchers coming from fields with overlap to Image Processing (Geoinformatics, Photogrammetry, 3D Vision, Remote Sensing).
We aim for dividing the workshop time between one or two keynote presentation from invited speakers, and oral presentations of the accepted papers.
We will encourage prospective authors to use publicly available data, share their code, and ensure that the results that they will publish can be reproducible.
Christophoros NIKOU received the Diploma in electrical engineering from the Aristotle University of Thessaloniki, Greece, in 1994 and the DEA and Ph.D. degrees in image processing and computer vision from Louis Pasteur University, Strasbourg, France, in 1995 and 1999, respectively. He was a Senior Researcher with the Department of Informatics, Aristotle University of Thessaloniki in 2001. From 2002 to 2004, he was a Research Engineer and Project Manager with Compucon S.A., Thessaloniki, Greece. He was Lecturer (2004-2009), Assistant Professor (2009-2013), and Associate Professor (2013-2018) with the Department of Computer Science and Engineering, University of Ioannina, Ioannina, Greece, where he has been a Professor, since 2018. During the academic year 2015-2016 he has been a Visiting Associate Professor at the Department of Computer Science, University of Houston, USA. Prof. Nikou was the General Chair of IEEE International Conference on Image Processing 2018 (ICIP’18). Since 2019 he is a member of the Conferences Board of the Signal Processing Society of IEEE. He is also an Associate Editor for IEEE Transactions on Image Processing. His research interests mainly include computer vision, pattern recognition, image processing and analysis and their application to medical imaging. He is a member of EURASIP and an IEEE Senior Member.
Lazaros GRAMMATIKOPOULOS is an Associate Professor at the Department of Surveying and Geoinformatics Engineering at the University of West Attica (UNIWA, Greece). He received his PhD in Photogrammetry and Computer Vision from NTUA in 2007 and has been teaching courses on Photogrammetry and Computer Vision, 3D Processing and Visualization, UAVs & Mobile Mapping since then. He has published 44 scientific articles in international journals and conferences and has been cited 836 times according to Google Scholar. Dr. Grammatikopoulos has also participated in 18 national and European Research Programs. His research interests include close-range Photogrammetry, SfM and SLAM, 3D segmentation, camera-to-Lidar calibration, texture mapping of 3D models, aerotriangulation, and single image reconstruction.
Eleni PETSA is a Surveying Engineer with a PhD in Photogrammetry from the National Technical University of Athens (1996). She is a Professor of Photogrammetry at the Department of Surveying & Geoinformatics Engineering of the University of West Attica (UNIWA, Athens, Greece), where she has been teaching graduate and postgraduate courses on Photogrammetry and Computer Vision. She has 65 publications in international scientific journals and conferences, which have been cited 890 times. She has also participated in 14 research programs. Her research interests are in photogrammetry and 3D computer vision, mainly focused on camera calibration, texture mapping of 3D models and single image reconstruction.
Giorgos SFIKAS received his B.Sc. and M.Sc. degrees in Computer Science from the Department of Computer Science, University of Ioannina, Greece in 2004 and 2007, respectively, and his Ph.D. degree in Image Processing and Computer Vision from the University of Strasbourg, France in 2012. He also holds a BA degree in History and Archaeology from the University of Ioannina. In 2014-2018, he worked as a Research Associate at the Institute of Informatics and Telecommunications of the National Center for Scientific Research “Demokritos” in Athens, Greece. During 2016-2020 he has worked as a visiting lecturer at the University of Ioannina and in 2018-2020 as a Research Associate at the Information Technologies Institute of the Centre for Research and Technology - Hellas (CERTH) in Ioannina, Greece. Today he is an Assistant Professor at the department of Surveying and Geoinformatics Engineering of the University of West Attica. His research interests include Machine Learning and Computer Vision. He has published more than 50 papers on journals and conferences on these fields (>1000 citations, h-index=15), including on top venues such as IEEE/CVF CVPR and ICML.
George RETSINAS received the Diploma degree in Electrical and Computer Engineering from National Technical University of Athens in February 2014. His diploma thesis, supervised by Prof. Petros Maragos, tackled fingerspelling, a subtask of sign language recognition. He received his Ph.D. degree from the Department of Electrical and Computer Engineering of National Technical University of Athens in February 2020. Part of his Ph.D. (2015-2018) was done in collaboration with the Institute of Informatics and Telecommunications of the National Center for Scientific Research “Demokritos”. The main research direction during this collaboration was document analysis and recognition, with emphasis on machine learning approaches. Additional research directions, addressed during his PhD, included bio-signal processing and compression of deep neural networks. Today he is a Post-Doctoral Researcher, focusing on vision applications in robotics, including 3D vision applications, in National Technical University of Athens, collaborating with University of Ioannina and Athena Research Center. He has published more than 35 papers on journals and conferences on these fields (>370 citations, h-index=10), including top venues such as IEEE/CVF CVPR and IEEE TPAMI.
Panagiotis DIMITRAKOPOULOS received his M.Eng. and M.Sc. degrees in Computer Science from the Department of Computer Science and Engineering, University of Ioannina, Greece in 2019 and 2020, respectively. His M.Eng and MSc theses focused on Deep learning detection, classification models and Bayesian Variational models, both applied on the Medical Imaging domain. He is currently a Ph.D. candidate student in the same department from 2020. His thesis is focused on the combination of Bayesian and deep learning methods for computer vision tasks. His research interests lie on Bayesian methods, Machine/Deep Learning, Computer Vision, Instance Segmentation/Detection, Medical Imaging.
Andreas El Saer holds a MSc Degree in Computer Science, and he is a PhD candidate in Computer Vision & Deep Learning all at the University of West Attica, Greece. He has great working experience in Photogrammetry, Computer Vision, and Geospatial domains as well as hands on experience within H2020 projects, in Deep Learning and software engineering. His previous experience consists of multi-disciplinary tasks ranging to many different domains which creates a strong and solid knowledge base that facilitates the exploration and exploitation of many disruptive technologies and novel ideas. He also has experience (participating in more than 10 H2020 and NSRF projects) in proposal writing and managing. His main research interests include Computer Vision, Photogrammetry, Deep Learning, Geospatial and AR & XR.