The Object Tracking Study Based on the Gabor Wavelet Character

This paper introduces a method of using the Gabor wavelet characters to track an object. Characters with Gabor wavelet characters vector describing the object´s edge characters are distilled. The LM algorithm is adopted when the count energy function is minimised. The object of the sequence image is matched with the image character template composed of Gabor wavelet characters. The Characters Template is renewed using Greedy algorithm when the moving object is tracked. When the moving object has some abnormal phenomena, such as translation, some scale changes, local shading, even fast background brightness changes, this method can still track the object