Content based Image Retrieval using Texture and Color Extraction based Binary Tree Structure

There are different methods of image retrieval where the meta-data is associated with the image, commonly called as keywords. Content based image retrieval is important research field in many applications. In this paper the CBIR system is proposed which introduces a new binary tree approach along with color and texture common in most of the CBIR system for finding similar images from the database to a given query image. There are different features of an image such as color, texture, shape, orientation, etc. In the proposed system color and texture are used as basic features to describe all the images. In addition, a binary tree structure is used to describe higher level features of an image. To extract color information, two histograms i.e. Hue and saturation of the image are used. And to extract texture information image quantization and wavelet decomposition is applied to each image blocks. The Hue is quantized into 360 levels and the saturation into 100 levels The binary tree structure is implemented based on steps provided .In this system, the feature extraction and wavelet decomposition for texture extraction is used to compute the feature vectors of any image which helps in retrieval process. This approach combines the color and texture features and binary partitioning tree method in order to find the images similar to a specific query image. The Minkowski difference equation is used to measure the distance. The image processing toolbox is available in the Matlab which consists of various inbuilt function to perform various operation on the image easily, which are difficult if they are implement using user defined function. The proposed system is implemented using the functions of Matlab software.