Dr. Petia Radeva is a full professor at the Universitat de Barcelona (UB), Head of the Consolidated Research Group “Artificial Intelligence and Biomedical Applications (AIBA)” at the University of Barcelona. She was PI of UB in 10 European and international and more than 35 national projects devoted to applying Computer Vision and Machine learning for real problems like food intake monitoring (e.g. for patients with kidney transplants and for older people). She is an Associate editor in Chief of Pattern Recognition journal (Q1, IP=7.196) and the International Journal of Visual Communication and Image Representation (Q2, IP=3.13). She is a Research Manager of the State Agency of Research (Agencia Estatal de Investigacio ́n, AEI) of the Ministry of Science and Innovation of Spain. Petia Radeva belongs to the top 2% of the World ranking of scientists with the major impact in the field of TIC according to the citations indicators of the popular ranking of Stanford. Also, she was selected in the first 6% of the ranking of Spanish and foreign most cited female researchers from any field according to the Ranking of CSIC. Moreover, she was awarded IAPR Fellow in 2015, ICREA Academia’2015 and ICREA Academia’2022 assigned to the 30 best scientists in Catalonia for her scientific merits, and received several international and national awards (“Aurora Pons Porrata” of CIARP, Prize “Antonio Caparro ́s” for the best technology transfer at UB, etc). She supervised 24 PhD students and published more than 100 SCI journal publications and 250 international chapters and proceedings, her Google scholar h-index is 53 with more than 11000 cites. Petia Radeva has published in more than 20 journals and international peer-reviewed conferences on food image analysis, She led IP 5 European and national projects related to the application of deep learning for non-intrusive food intake monitoring.
Title: Data-centric Food Computing
Abstract: Deep Learning (DL) has made remarkable progress, achieving super-human performance. However, when it comes to classifying a complex domain as food recognition, there is still much room for improvement. Additionally, DL relies on greedy methods that require thousands of annotated images, which can be a time-consuming and tedious process. To address these issues, we will discuss several data-centric approaches that help to the problem, specially how self-supervised learning offers an efficient way to leverage a large amount of non-annotated images and to make DL models more robust and accurate. Moreover, we will present how a new combination of self-supervised, and prompt learning can help to the fine-grained food recognition.
Dr. Chong-Wah Ngo is a professor with School of Computing and Information Systems, Singapore Management University (SMU), Singapore. He received his PhD in Computer Science from Hong Kong University of Science & Technology. Before joining SMU, he was a professor with the Department of Computer Science, City University of Hong Kong, Hong Kong. His main research interests include multimedia search, multi-modal fusion, and video content understanding. Recently, he is active in lifestyle and wellness research from the perspective of artificial intelligence, where his research includes food and portion size recognition, risk analysis and lifestyle behavior nudging. He has researched on various issues in food computing, including ingredient recognition, open-vocabulary ingredient segmentation, food image generation from recipe, causality based cross-domain food recognition, cross-modal and cross-lingual food image and recipe retrieval, mobile app food logging systems. Currently, he leads the research and development of FoodAI engine in SMU, an engine for recognizing local Singapore food and has been deployed by Singapore Health Promotion Board for citizen food logging. His research results in food computing are published in CVPR, ACM Multimedia, and AAAI conferences. Chong-Wah Ngo is currently the associate editor of ACM Trans. on Multimedia Computing, Applications, and Communications. He is the general chair of The Web Conference (WWW) 2024 and program chair of ACM Multimedia 2019. He served as the chairman of ACM (Hong Kong Chapter) during 2008-2009 and was named as ACM Distinguished Scientist in year 2016.
Title: Large Multimodal Model for Food Recognition
Abstract: If we are what we eat, the history of food consumption could be the key determinant for the prediction of future health. However, food logging in the free-living environment is known as a highly challenging problem. In this talk, I will discuss the basic building blocks towards the holy grail of developing a smart agent that can quantify food consumption and provide dietary advice. A review of the recent progress in various tasks relevant to food recognition will be summarised. These include ingredient recognition, segmentation, counting and portion size estimation. I will then present the current efforts in building a large multimodal model (LMM) to unify these tasks for food conversation about dietary planning and suggestion. The limitations of LMM and the potential of extending LMM for healthcare applications will be discussed.