We present a new type of local image descriptor which yields binary patterns from small\nimage patches. For the application to fingerprint liveness detection, we achieve rotation\ninvariant image patches by taking the fingerprint segmentation and orientation field into\naccount. We compute the discrete cosine transform (DCT) for these rotation invariant\npatches and attain binary patterns by comparing pairs of two DCT coefficients. These patterns\nare summarized into one or more histograms per image. Each histogram comprises\nthe relative frequencies of pattern occurrences. Multiple histograms are concatenated and\nthe resulting feature vector is used for image classification. We name this novel type of\ndescriptor convolution comparison pattern (CCP). Experimental results show the usefulness\nof the proposed CCP descriptor for fingerprint liveness detection. CCP outperforms\nother local image descriptors such as LBP, LPQ and WLD on the LivDet 2013 benchmark.\nThe CCP descriptor is a general type of local image descriptor which we expect to prove\nuseful in areas beyond fingerprint liveness detection such as biological and medical image\nprocessing, texture recognition, face recognition and iris recognition, liveness detection for\nface and iris images, and machine vision for surface inspection and material classification.
Loading....