Main navigation
- Programs
- Subjects
- Universities
- Destinations
- Advice
Our research focuses on deciphering visual perception to develop machines with enhanced visual intelligence. This is achieved through core principles in reasoning, prediction, and various learning approaches including supervised, semi-supervised, and unsupervised methods, along with stochastic optimization. Our work spans (i) foundational computer vision areas like video analysis, human pose estimation, object recognition and tracking, (ii) practical applications such as medical imaging, interpreting social behaviors from video interactions, autonomous vehicle collision systems, and other video-based regression challenges, and (iii) the convergence of computer vision and graphics, where we create lifelike avatars for more natural human interaction. We continuously explore novel applications of computer vision, including cardiac 3D modeling from MRI scans, spotting deception through visual signals, analyzing group dynamics in sports like volleyball, enhancing STEM education via video analysis, and similar cutting-edge challenges.