Content based image retrieval using feature extracted from dot diffusion block truncation coding

Identification of Similar Images from a large image database a critical issue in the Image Processing. For the purpose of retrieving more similar image from the large image database a new technique is proposed in this Paper. To derive the image feature descriptor Dot Diffusion Block Truncation Coding (DDBTC) is employed. The image feature descriptors are simply derived from two DDBTC color quantizers and its corresponding Bitmap image. Color Histogram Features (CHF), Color Co-occurrence Features (CCF) are derived from the two color quantizers whereas Bit Pattern Feature (BPF) is derived from Bitmap image. The color quantizer represents the global characteristics while the Bitmap image represents the local characteristics of the image. Similarity between two images can be measured in terms of different distance metrics. Experimental result shows that the superiority of the proposed technique in terms of Average Precision rate (APR) and Average Recall Rate (ARR) under Natural images.