Tracking in Low Frame Rate Video: A Cascade Particle Filter with Discriminative Observers of Different Lifespans

Tracking in low frame rate video is a practical problem to solve. This paper presents a temporal probabilistic combination of discriminative observers with different lifespans, where the observers are learned from different ranges of samples, with different subsets of features, to ensure a complementary fusion. The paper is by Yuan Li, Haizhou Ai, Takayoshi Yamashita, Shihong Lao and Masato Kamade, and it received the CVPR 2007 best student paper award. Here (PPT) is a presentation by Qi Zhao. Fall 2007.