Image retrieval using local and global properties of image regions with relevance feedback

Abstract

This paper proposes an image retrieval system using the local and global properties of image regions. Colour features are extracted using the histograms of HSV colour space, texture features using Gray level Co-occurrence matrix (GLCM) and shape features using Edge Histogram Descriptors (EHD). The object regions are roughly identified by segmenting the image into fixed partitions and finding the white pixel density in each partition using edge thresholding and morphological dilation. To improve the retrieval efficiency, global colour and shape features are also taken into account. Euclidean distance measure is used for computing the distance between the features of the query and target image. An automatic relevance feedback algorithm is also proposed for improving the retrieval accuracy. Preliminary experimental results show that the proposed method provides better retrieving result than retrieval using some of the existing methods.