A Novel Image Semantic Block Clustering Method based on Artificial Visual Cortical Responding Model

This paper proposed a novel visual information process model named artificial visual cortical responding model (AVCRM) to obtain the invariable time sequence response feature from the sub-image block in the image. By compressing the time sequence feature and selecting the important points in the sequence, we compared the compressed version of sequences with each other to generate the distance matrix. According to distance matrix, we clustered the sub-images into the initially manually assigned concept categories to attain the semantic distribution map of the image. This mechanism was proved to be effective through the experiments and made a good semantic foundation of the future content based image retrieval research work