跳到主要內容區塊

Visual AI for Digital Health: Opportunities & Challenges

 

Talk title

Visual AI for Digital Health: Opportunities & Challenges

 

Abstract

Visual AI refers to the ability for computer to see images like what a human would. Just as when users are familiar with visual capabilities of their smartphones for social and work communications, they could also tap on such capabilities for interesting digital health applications.  Unfortunately, these applications still require significant research advances in visual AI to address new challenges. Using Food Visual AI (FVAI) capabilities for illustration, this talk will present how food image recognition and segmentation have found their use in several interesting digital health applications. While the computer vision research community has made substantial progress in recent years, there are still many important challenges in FVAI that remain to be addressed in order for the research works to make real impact to people's health.  We will cover these challenges and how our research team overcame some of them.  In particular, we will describe our research endeavour in FoodAI and FoodAI+ projects. The talk will end with a few suggestions of what researchers can do to collectively change the current FVAI and digital health research.

 

About the Speaker

Dr. Ee-Peng Lim is the Lee Kong Chian Professor of Computer Science with the School of Computing and Information Systems at the Singapore Management University (SMU). He was also the Director of Living Analytics Research Centre in SMU, a National Research Foundation (NRF) supported research centre focusing developing personalized and participatory analytics capabilities for smart city and smart nation relevant applications. Dr Lim received his PhD degree from University of Minnesota. His research expertise covers social media mining, social and urban data analytics, and information retrieval. He is the recipient of the Distinguished Contribution Award at the 2019 Pacific Asia Conference on Knowledge Discovery and Data Mining (PAKDD), and the Test of Time awards at 2020 ACM Conference on Web Search and Data Mining (WSDM) and 2021 European Conference on Information Retrieval (ECIR).  More