Image feature fusion and its application based on trace transform and improved GLBP

Abstract

In order to solve the problem that the features extracted by the trace transform are not good enough for the description of face images, a target recognition method based on the image transform and the improved gradient local binary pattern (GLBP) is proposed. This method combines the trace transform and the improved gradient local binary mode. By selecting sampling points on the trace line and carrying out GLBP coding, the GLBP texture feature information that can describe the whole trace line is extracted. Classification experiments on ORL face database show that under the circumstances of less training samples, the target recognition based on trace transform and GLBP method to extract the texture characteristics, its recognition ability has obvious improvement than trace transform, within a smaller fluctuation range, better stability and better resolution for texture image.