Local Extreme Co-occurrence Edge Binary Pattern for Bio-medical Image Retrieval

A new approach to bio-medical image retrieval, local extreme co-occurrence edge binary pattern (LECoEBP) pattern proposed. Local extreme edge binary patterns (LEEBP) give the texture information and gray level co-occurrence matrix finds the co-occurrence of binary patterns. LEEBP extracts the four edges for each pixel and GLCM applied on each edge to find the joint occurrence of edge binary patterns. In LEEBP, each edge has directional relative information of all possible directions of a pixel. The directional relative information calculated based on the sign code magnitudes of local differences from center pixel to its directional pixels. The combination of LEEBP and GLCM showed significant results compared with the current algorithms on benchmark databases Emphysema-CT and OASIS-MRI.