The interplay of machine learning, image understanding and video analysis
Li Cheng (University of Alberta, Canada)
NICTA SML SEMINARDATE: 2005-11-15
TIME: 10:00:00 - 11:00:00
LOCATION: RSISE Seminar Room, ground floor, building 115, cnr. North and Daley Roads, ANU
CONTACT: JavaScript must be enabled to display this email address.
ABSTRACT:
Recently, vast amounts of image and video data have become available due to increases in storage and advances in hardware technology. The abundance of data has created significant new challenges and opportunities for computer vision research. I am particularly interested in improving techniques for understanding images and video in order to replace tedious human labor in concrete tasks. My approach is to apply machine learning techniques to automate the development of computer vision systems, exploiting the substantial data resources that have become available.
In this talk, I will present my work on image understanding using graphical models and Bayesian methods. Specifically, I will talk about a Bayesian approach to Stereo Vision. The goal of stereo vision is to infer the optimal disparity map for a given pair of images. A Bayesian framework is proposed where solutions to both the model parameters and the disparity map are unified as predictions of latent variables. A mixture of sampling and deterministic inference is adopted to balance between effectiveness and efficiency: the parameters are estimated via Markov Chain Monte Carlo sampling techniques and the Maximum A Posteriori (MAP) disparity map is approximately inferred via a deterministic algorithm.
Additionally, I will present my current work on realtime video analysis,
which addresses the problem of foreground background separation using
online learning with sparse kernels, and show some experimental results.
BIO:
Li Cheng received the PhD degree in Department of Computing Science from University of Alberta, Canada in 2004. Previously, he received the ME degree from Nankai University, and the BS degree from Jilin University, China. Currently he is a research associate in the same department of University of Alberta. His research interests are mainly on Machine Learning, Image Understanding and Video analysis.
