Content Based Image Mining through Correlative Histogram Feature Selection

In this paper, content-based image retrieval (CBIR) for content-based image mining is proposed. Here, an approach of image detection in a image dataset and an improved histogram based approach for Correlative Histogram Features named as Correlative Histogram based Coding (CR-HBC) is proposed. The proposed approach enhances the objective of processing noise and system overhead with respect to the representation and retrieval performance. The proposed approach executes an inter-class searching mechanism to localize the object model by discarding noise elements. Further, a histogram similarity model and a spectral domain feature representation are proposed. The experiments have been conducted on the benchmark datasets. The experimental results show that the processing resource overhead and retrieval efficiency are improved.