I am interested in developing practical real-time computer vision systems to solve real-world problems.
My primary research focus is in Action/Activity/Gesture recognition in videos taken from YouTube (Consumer Videos), aerial platforms (Aircraft), rooftop surveillance cameras, and Kinect Sensor with depth information.
Few interesting projects that I worked on during my Ph.D
- Worked on Video Image Retrieval and Analysis Tool (VIRAT) project for three years and led UCF computer vision group in VIRAT Phase II.
- Partly involved in PASCAL challenge, KINECT Gesture challenge and Automated Low-Level Analysis and Description of Diverse Intelligence Video (ALADDIN) project.
- Developed two challenging action recognition datasets “UCF50” and “UCF-Aerial-Rooftop-Ground” (UCF-ARG) and released for public use.
- Built action recognition system “KEYS” and presented a live demo at ICCV 2009 in Japan.
- Worked and led a team on an NIH project to detect and segment brain tumors in MRI images.