Dynamic Texture Feature Extraction Using Weber Local Descriptor

摘要

In numerous biometric applications, gender recognition from facial pictures plays a vital role. During this paper, we have a tendency to investigate Weber's native Descriptor (WLD) for gender recognition. WLD may be a texture descriptor that performs higher than alternative similar descriptors however it's holistic as a result of it's terribly construction. We have a tendency to extend it by introducing native abstraction information; divide a picture into variety of blocks, calculate WLD descriptor for every block and concatenate them. The gender from face options are classified by the minimum distance measurement and neural network.