Incremental discriminative color object tracking

Abstract

This paper presents an object tracking algorithm based on discriminative 3D joint RGB histograms of the object and surrounding background. Mean-shift algorithm on the object confident map is used for localization. An incremental color learning scheme with a forgetting factor is utilized to account for appearance variation of the object. Evaluated against three state of the art methods, experiments demonstrate the effectiveness of the proposed tracking algorithm where the object undergoes variation in illumination and color. Implemented in MATLAB, the proposed tracker runs at 25.7 frames per second.

Tasks

Reproductions