My research spans four related areas of vision, namely, machine vision, image processing, visual perception, and visual neuroscience. This includes the development of reliable algorithms for machine vision systems, and theoretical work on the neural basis of visual perception in biological systems.

Most of my work to date has focused on visual motion analysis and binocular stereopsis. Understanding these visual processes is central to the development of algorithms that would, for example, facilitate the determination of 3d scene layout, the detection and tracking of objects, the inference of the 3d motion of an observer (camera), and motion-based recognition of objects and their activities. Potential applications include novel human-computer interfaces, computer graphics and special effects, assisted automotive navigation, security and surveillance, and human motion capture.

Computer Vision and Image Processing

3D Hand Pose Tracking

Physics-Based People Tracking

Gaussian Process Latent Variable Models for Human Pose and Motion

3D People Tracking

Motion Layers and Image Parsing

Appearance Models and 2D Tracking

Appearance Changes in Image Sequences

Motion (Occlusion) Boundaries

Parameterized Models for Optical Flow

Phase-Based Methods (Optical Flow & Binocular Disparity)

Evaluation and Comparison of Optical Flow Techniques

Spatiotemporal Filters for Optical Flow

Embedding Imperceptible Signals in Images