Center for Computer Vision

Director: Prof. Hongbin Zha

Overview:

The Center for Computer Vision is dedicated to the investigation of "Object Recognition, Scene Understanding, and Behavior Analytics in Real Environments." The primary objective is to leverage advanced sensing and imaging technologies to develop robust and efficient visual perception theories and methods for intelligent machines. The center focuses on the advancement of multi-scale scene dynamic 3D reconstruction techniques to address the challenges posed by structurally complex and dynamically changing real environments and large-scale scenes. This research provides a solid technical foundation for real-time human-machine-environment interactions.

Additionally, we conduct extensive research in the fields of object, scene, and event recognition, aiming to enhance the adaptability of visual systems to various environments through approaches such as online learning.

Main Research Directions:

• Statistical computational models for visual understanding

• Image enhancement, attribute understanding, and object recognition

• Semantic segmentation and analysis of images and videos

• Scene 3D dynamic reconstruction and interpretation

• Behavior recognition and intent understanding