A Comparative Study on CBIR Using Color Features and Different Distance Method

Image retrieval system has been effectively and efficiently used tool for large image databases. In CBIR system, the input is query image and output is retrieved similar images. CBIR is the method to retrieve images on visual content. Several retrieval approaches have been developed to increase the performance of CBIR system. In this paper, we present a comparative analysis of CBIR using color features. We also discuss about color features such as Color statistics, Color moment, Dominant color descriptor (DCD), HSV histogram, Color Correlogram, and Color String. In this research, we find distance between two images using different similarity metrics, namely Jaccard distance (JD), Hamming distance (HD), Euclidean distance (ED), Cheby-Chev (CD), and Manhattan distance (MD). Support Vector Machine (SVM) is used for classification of categories. The comparative results on Corel database with 1000 images. The experimental result is on precision for calculating system performances.