KERNEL SELECTION BY MUTUAL INFORMATION FOR NONPARAMETRIC OBJECT TRACKING
J.M. Berthomme, T. Chateau, Michel Dhome · 2012
This paper presents a novel method of selecting kernels for the subsampling of nonparametric models used in real-time object tracking in a video stream. We propose a method based on mutual information, inspired by the CMIM algorithm (Fleuret, 2004) for the selection of binary features. This builds, incrementally, a model of appearance of the object to follow, consisting of representative and independant kernels taken from points of that object. Experiments show gains, in terms of accuracy, compared to other sampling strategies.