An Effective Content Based Image Retrieval Using Dot Diffusion Block Truncation Coding

摘要

Image processing applications such as medical diagnosis, crime prevention, publishing or advertising, utilizes the traditional image processing techniques. Simple browsing can easily identify images from a small collection of images but for large and different collection of images identification of images emerge as a critical issue. In this paper for image retrieval four different features are extracted using DDBTC technique. The first two features such as Color Co-occurrence Features (CCF) and Color Histogram Features (CHF) are obtained using color quantizers, Bit Pattern Feature (BPF) and Bit Histogram Feature (BHF) are obtained using Bitmap image. In order to remove the false counter problem and blocking effect different error diffusion kernels are employed. The results shows better accuracy.