NAVER LABS Europe seminars are open to the public. This seminar is virtual and requires registration
Date: 23rd March 2023, 10:00 am (CET)
Relational and structural vision with high-order feature transforms
About the speaker: Minsu Cho is a Visiting Faculty Researcher at Google Research and an Associate Professor at POSTECH, South Korea, leading POSTECH Computer Vision Lab. Before joining POSTECH in the fall of 2016, he worked as a postdoc and a starting researcher at Inria WILLOW team and École Normale Supérieure, Paris, France. He completed his Ph.D. in 2012 at Seoul National University, Korea. His research lies in the areas of computer vision and machine learning, especially in the problems of visual semantic correspondence, symmetry analysis, object discovery, action recognition, and minimally-supervised learning. He is interested in the relationship between correspondence, symmetry, and supervision in visual learning. He is an editorial board member of the International Journal of Computer Vision (IJCV) and has been serving as an area chair in top conferences including CVPR, ICCV, and NeurIPS.
In 2020, he was inducted into the Young Korean Academy of Science and Technology (Y-KAST).
Abstract: Our visual world is full of complex patterns with multiple constituents and thus visual understanding often requires the ability to grasp the relations and structures of visual elements, i.e., which visual elements are related to each other and how visual elements are structured as a whole. Despite the remarkable progress, modern neural networks for computer vision are largely limited in the relational and structural awareness of visual patterns, which contrasts starkly with human vision. In this talk, I will present a series of my recent work on high-order feature transforms that address the aforementioned limitations and enable minimally-supervised recognition and structural perception of images and videos, focusing on few-shot classification/segmentation and motion-aware action recognition.