An Efficient Content Based Image Retrieval Using Block Color Histogram and Color Co-occurrence Matrix

摘要

In the present days, Images are widely used everywhere and the image retrieval plays a vital role. Content Based Image Retrieval (CBIR) system is a well-known technique for effective image retrieval. The Proposed CBIR system use color and spatial feature to retrieve similar images. Color histogram is the widely used method to extract color features which liberate translation, rotation and scaling of image and avoid spatial feature. The color co-occurrence matrix extracts spatial feature. The Proposed CBIR system uses 3*3 block color histogram with 1:2:1 proportion of segmentation with weight and color co-occurrence matrix to extract color and spatial feature. The images in the database are indexed with feature vectors by using SR-tree algorithm which is used to increase the speed of retrieval. The feature matching process is carried out using Euclidean distance of color histogram and color co-occurrence matrix and find distance metric. Then the images in the database are sorted using distance metric and required number of images will be retrieved.