|
|
|
My Research Programme
Many real world vision tasks such as motion segmentation, large scale scene reconstruction, and object/scene classification remain challenging research problems. This is not surprising, considering that more than half of the neocortex is involved in visual processing. To solve these vision problems robustly, we envisage that it requires a concerted research effort in integrating the different vision modules together, as well as further advancing our fundamental understanding of the individual vision modules.
I am interested to investigate the feedback and lateral links that exist in complex vision problems. For instance, the segmentation problem will need to combine a broad range of technical advances in computation of 3D surfaces, knowledge about natural scene statistics and gestalt laws, and expertise in advanced mathematical techniques such as level set, graph cut, MRF-based learning. It provides an opportunity for psychophysical, computational and mathematical studies that will extend our understanding of the processes underlying the interactions between various vision modules, as well as the incorporation of environmental constraints to enhance the performance in dynamic, real-world situations.
In the past decade, my research has revolved around the central question of space perception arising from motion cues, otherwise known as the structure from motion (SFM) problem. I approach this problem mainly from the computational perspective but also study the psychophysical implication and some applied aspects (see the links in the next few section). It remains an active area of my research. Now, I am using sparsity-based techniques to address various motion problems such as 3D motion segmentation, as well as change detection amidst scenes with complicated dynamic behavior such as swaying trees and undulating waves.
For prospective research students, I am looking for someone who is really keen to understand the visual processes involved in human vision, and feels excited to build a robust system that can function in the real world. Preferably, the student must have an adequate level of mathematical sophistication, as the field of computer vision is currently going through a crucial mutation, requiring more and more mathematical skills such as PDE, differential geometry, functional analysis, etc.. Interested students with EE, CS, or applied mathematics background are welcome to contact me.
Current Research Topics
2. Visual Psychophysics in
motion and stereo
3. Applied Motion-based Video Indexing
SAMPLE PUBLICATION
Computational Domain
Geometry of Distorted Visual Space . Int'l Journal
of Computer Vision, 32(3), pp 195-212, 1999. © 1999 by Kluwer academic
Error in Depth Reconstruction . Int'l Journal of
Computer Vision, 44(3), pp 199-217, Aug 2001. © 2001 by Kluwer academic
Behaviour of SFM algorithms . Int'l Journal of
Computer Vision, 51(2), 111-137, 2003. ©
2003 Kluwer academic
Depth distortion under calibration uncertainty, Computer Vision and Image Understanding, Volume 93, Issue 3 , March 2004, Pages 221-244.
How do we Perceive Depths from Motion Cues in the Movies: A Computational Account, Journal of the Optical Society of America A: Optics, Image Science, and Vision, 2008.
Linear Quasi-Parallax SfM using Laterally-placed Eyes, International Journal of Computer Vision, Volume 84, Number 1 / August, 2009, pg 21-39.
Hierarchical Spatio-Temporal Context Modeling for Action Recognition, CVPR09.
When Discrete Meets
Differential — Assessing the Stability of Structure from Small Motion,
International Journal of Computer Vision, vol. 86, nos 1, pp. 87-110, 2010.
Error
Characteristics of SFM with Erroneous Focal Length, Computer Vision and
Image Understanding, 115, No.1, (Jan 2011) 16–30.
Smoothly
Varying Affine Stitching, CVPR 2011, Jun 20-25, Oral presentation
Active Visual Segmentation, IEEE Transaction on Pattern Analysis and Machine Intelligence (TPAMI), Vol 34, No. 2, p639-653, April 2012.
Simultaneous Camera Pose and Correspondence
Estimation with Motion Coherence,
International Journal of Computer Vision, 96(2): 145-161 2012
Quasi-Parallax
for Nearly Parallel Frontal Eyes --a possible role of binocular overlap during
rapid locomotion, accepted for publication in International Journal of
Computer Vision, June 2012.
Seeing
Double Without Confusion: Structure-from-Motion in Highly Ambiguous Scenes,
CVPR 2012.
Block-sparse
RPCA for Consistent Foreground Detection, accepted for ECCV2012.
Psychophysical Domain
Slant and Tilt Perception: A computational and psychophysical study . “Lecture Notes in Computer Science” Vol 1843, 2000. © 2000 by Springer-Verlag
Absolute distance perception during in-depth head
movement: Calibrating optic flow with extra-retinal information.
Vision Research, 42(16), pp. 1991-2003, 2002.
The visual perception of plane tilt from motion in small field and large field: psychophysics and theory, Vision Research, Volume 46, Issue 20, October 2006, Pages 3494-3513.
Others
Scene-based shot change detection
. Computer Vision and Image Understanding, 79: (2) 224-235 Aug 2000.
© 2000 by ACADEMIC PRESS
Establishment
Shot Detection Using Qualitative Motion. IEEE Conference on Computer Vision
and Pattern Recognition, June 18 - 20, 2003,
Framework
for synthesizing semantic-level indexes. Multimedia Tools and Applications
20(2): 135-158; Jun 2003.
Synergizing Spatial and
Temporal Texture. IEEE Transactions on Image Processing 11(10), pp.
1179-1191, 2002.
Addressing the problems of
Bayesian Network classification, IEEE Transactions on Knowledge and Data
Engineering, Volume 16 , Issue 2, February 2004, Pages: 230 – 244.
Affective Understanding in Film , IEEE Transactions on Circuits and Systems for Video Technology, Volume: 16 Issue: 6 June 2006. Page(s): 689- 704.
A Taxonomy of Directing Semantics for Film Shot Classification, IEEE Transactions on Circuits and Systems for Video Technology, Volume 19, No. 10, October 2009, pp. 1529-1542.
My Teaching
· EE6901 3D Vision & EE4212 Computer Vision, the former a Ph.D. level course and the latter a fourth year course.
· Also leading a seminar class EE6903 Advanced Models Of Biological Perception and EE6733 Advanced Topics on Vision and Machine Learning.