Download now Free registration required
LBP is a very powerful method to describe the texture and shape of a digital image. Therefore, it appeared to be suitable for feature extraction in face recognition systems. A face image is first divided into small regions from which LBP histograms are extracted and then concatenated into a single feature vector. This feature vector forms an efficient representation of the face and can be used to measure similarities between images. The Local Binary Pattern (LBP) was originally designed for texture description. It is invariant to monotonic grey-scale transformations which is very important for texture analysis. Also due to the computational simplicity processing of image in real time is possible.
- Format: PDF
- Size: 527.8 KB