Face Recognition with Name Using Local Weber‟s Law Descriptor

In Image processing face recognition plays an important role in various biometric applications. WLD (Weber’s Local Descriptor) will be used for face recognition. WLD is a texture descriptor that performs better than other similar descriptors but it is holistic due to its very construction. Image is divided into number of blocks and WLD is calculated for each block and then concatenate them. This spatial WLD has better discriminatory power compared to other existing descriptor methods. It is used to represent the image in terms of differential excitations and gradient orientation histogram for texture analysis. The WLD is based on Weber’s law and it is robust to illumination change in noise and other distortions. So it effectively analyzes the facial features for accurate matching with the existing facial images in the database. The feature extraction approach will be used for both test and database images to recognize face. The face will be recognized by finding Euclidean distance between them. The proposed spatial WLD with simplest classifier gives much better accuracy with lesser algorithmic complexity than other existing face recognition approaches. Through a large number of experiments performed on FERET (Facial Recognition Technology) database, two datasets with low (20x16) and high (60 x48) resolutions from FERET database are used.