Content Based Image Retrieval Using Boosted Local Extrema Co-occurrence Pattern

Authors

  • Naveen Kumar N
  • S. Rajkumar

Abstract

This paper presents the Boosted Local Extrema Co-occurrence Pattern (BLECoP) which focuses more on the local structures by enhancing the local extrema pattern. The proposed method can effectively combine with structural texture, GLCM and HSV color space. BLECoP tested over the standard Corel databases, which is extensively used in the Content-Based Image Retrieval(CBIR) domain. Experimental results show that BLECoP outperforms other state-of-the-art descriptors that are recently proposed in CBIR research using color and texture features. The performance of the proposed descriptor’s image retrieval results is achieved best accuracy rates, that are remarkable on heterogeneous(non-flat) images. Further BLECoP can be effectively used to retrieve more accurate images in a very large scale image databases.

Downloads

Published

2019-12-16

Issue

Section

Articles