Hybrid Face Anti-Spoofing Framework using Surface Texture Feature

There are various techniques for extracting features from images, such as Grayscale Pixel Value Features, Mean Pixel Value of Channels, Extracting Edge Features, Image Surface Texture Features, LBP and GLCM Surface Texture Features. Respective approaches are critical for computer vision (CV2) and image processing and classification. Among these LBP-based algorithms and their variations, GLCM-based algorithms and there variations are widely accepted and used just because of their simplicity, understanding and high efficiency. GLCM is very powerful algorithm which can tell every aspects of any image because this algorithm studies the detailed features of an image. However, the native LBP based algorithm calculates the significant difference between local centre of the LBP kernel and its adjacent pixel which is deffiend as neghibours of the central pixel and the magnitude difference is what called the surface texture of an image. Furuthere more to obtain detailed surface texture featuresof image without augmenting the feature dimensions of LBP objects, this paper proposes a nobel surface texture extraction method where the features are extracted with different LBP objects having other parameters and concatenated along the x-axis, e.g., (LBP1+LBP2+...). Furthermore, they are combined with GLCM features.