Title: Single-View 3D Scene Analysis

Reporter: Dr. Prof. James Elder, Department of Electrical Engineering & Computer Science at York University, Canada

Time:  9:30AM-11:00AM, 24 Oct. 2016

Address:3-202, Building No. 5, School of Remote Sensing and Information Engineering, Wuhan University



Computer vision systems are beginning to rival human vision for certain tasks. However, there are still many ways in which human and machine vision systems are divergent. While most single-view computer vision algorithms are 2D in nature, humans find it almost impossible to see in 2D: an innate 3D sense seems to provide the perceptual scaffolding for our understanding of the world around us.

A second divergence concerns the way we learn. Most computer vision learning is supervised, and the resulting systems, once trained, do not adapt. Human perception, on the other hand, appears to be extremely adaptive over many time scales.

In this talk, I will describe recent research in our lab on single-view scene analysis that attempts to embrace these principles of 3D adaptive perception. Specifically I will review how these principles have led to recent progress on the problems of online camera pose estimation, roadway analytics and crowd estimation.


Dr. James Elder is a Professor in the Department of Electrical Engineering & Computer Science and the Department of Psychology at York University, and a member of York’s Centre for Vision Research. His research seeks to improve machine vision systems through a better understanding of visual processing in biological systems. Dr. Elder’s current research is focused on natural scene statistics, perceptual organization, contour processing, shape perception, single-view 3D reconstruction, face detection, attentive vision systems and machine vision systems for dynamic 3D urban awareness.

Over his career Dr. Elder has spearheaded and led numerous successful collaborative research projects with funding from NSERC, OCE, GEOIDE, IRIS, DRDC, DND and MTO totaling more than $6M and currently leads an NSERC CREATE Training Program in Data Analytics & Visualization. He has published over 60 papers in leading international journals, conferences and books in both computer and biological vision, and these have been cited more than 3,800 times (Google Scholar). He also holds two patents on attentive vision technologies.

Dr. Elder’s research has won a number of awards and honours, including the Premier’s Research Excellence Award and the Young Investigator Award from the Canadian Image Processing and Pattern Recognition Society. He is appointed to the Editorial Boards of two major international journals.

Presented by Dr. Prof. James Elder