Local binary pattern: An improved LBP to extract nonuniform LBP patterns with Gabor filter to increase the rate of face similarity

• Published in: 2016 Second International Conference on Cognitive Computing and Information Processing (CCIP) pp. 1 - 5
• Main Authors: Y.B., Ravi Kumar, C.N., Ravi Kumar
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.08.2016
• Subjects: Data mining; Face; face descriptor; Face recognition; Facial features; Feature extraction; Gabor filter; Gabor filters; Histograms; Non-Uniform LBP Pattern
• DOI: 10.1109/CCIP.2016.7802878

The LBP is a technique used to extract textures of an image, especially in facial image analysis. As we are intended to extract more fine texture details of an image to increase the rate of facial similarity, we combined the techniques of LBP with Gabor filters to obtain a fine texture details. The Gabor filters are designed to multiply the filter coefficients with an LBP pattern to get more finely textured images. The gabor filter used in this proposed method helped to filter the coefficients of LBP pattern with gabor filters. The gabor filter is a technique used to generate a real and imaginary parts of a trignometric function with respect to θ. Similarly, LBP pattern is an individual technique used to extract a texture of an image in terms of local binary patterns. By combining the two individual techniques into a single technique and by employing a non-uniform LBP Pattern, we will be able to extract more finely textured images of a dataset. The result of combining LBP with Gabor filter produces a result of 93.4%, which is better than individual approaches.