Articulated Pose
Estimation Theory for CV
Ph.D. Research
Contact Details
Awards and Grants
Community Duties

Ph.D. Research

Disparity Map Completion for Trilinear-Tensor View Synthesis
from Wide-Baseline Stereo

In my Ph.D., I proposed a novel human-model-based disparity map completion method for high-quality novel-view synthesis from wide-baseline stereo, using trilinear-tensor transfer. Using prior knowledge in the form of an articulated human body model, I developed a method for disparity map completion which included novel contributions in articulated human body pose estimation and articulated deformable model fitting to an unstructured cloud of data points. The core novel contribution of the thesis was the disparity map completion in disparity space instead of 3-D, which involved setting the entire process, from articulated pose estimation to model fitting and disparity map generation, in disparity space using the principles of multi-view geometry, thereby completely avoiding the step of 3-D reconstruction and consequently significantly reducing the computational complexity of the method as well as increasing its accuracy.