B-Scan Images Analyzed By CNN and Co-Occerrence Matrix

In this paper, we combine cellular neural network (CNN) and gray step co-occurrence matrix to process B-scan images of fatty patients´ livers. We deal with the B-scan images of fatty patients´ livers by the edge detection cellular neural network, and then analyze the B-scan image features, including the co-occurrence matrix´s contrast (Contrast), correlation (Correlation), energy (Energy) and homogeneity (Homogeneity). The value of Contrast on 0deg direction seems to correlate to the degree of the damage of patients´ livers. It is expected that the method provided in this paper will be helpful to the diagnosis of biomedical images.