Feature Replenishment for Long-Term Visual Motion Tracking

This paper presents the feature replenishment technique used by our 3D Feature Based Tracker (3DFBT). The 3DFBT is a distributed, agent-based system for real-time 3D visual motion tracking using stereo. Feature replenishment means that the system can find new trackable features to replace those lost to noise and occlusion, and on new objects entering the scene, or revealed by dis-occlusion. This allows tracking to be continued over longer periods.

[1]  L. Kitchen,et al.  Issues, Architectures and Techniques in Real-Time Vision , 1992, Proceedings of the Third Annual Conference of AI, Simulation, and Planning in High Autonomy Systems 'Integrating Perception, Planning and Action'..

[2]  Dana H. Ballard,et al.  Animate Vision , 1991, Artif. Intell..

[3]  James W. Cooper,et al.  Early jump-out corner detectors , 1991, Proceedings. 1991 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[4]  L. Kitchen,et al.  A Distributed Systems Architecture for Real-time Computer Vision 1 Real-time Task-directed Robot Vision , 2022 .

[5]  Roger Y. Tsai,et al.  A versatile camera calibration technique for high-accuracy 3D machine vision metrology using off-the-shelf TV cameras and lenses , 1987, IEEE J. Robotics Autom..