Video Synopsis for Surveillance
Efficient keyframe extraction and dynamic video synopsis methods for condensing long surveillance videos.
Led a research team of 5 at Southeast University to develop video synopsis technology for surveillance video, focusing on efficient keyframe extraction and dynamic synopsis methods.
Key contributions:
- Implemented Graph-cut algorithms for optimization and classic clustering/motion-based methods for keyframe extraction in C++ and Python
- Proposed an efficient method based on motion-volume to extract keyframes, with performance comparable to state-of-the-art methods
- Completed the project ahead of schedule and achieved top 5% rank in project defense and evaluation
This was an undergraduate research project supervised by Prof. Siyu Xia.