Indexing and retrieval of images is a very important research topic that has attracted great interest in recent years. Image coding is a successful field that has been studied for over three decades. All image coding techniques attempt to visually extract the most important features and represent them as compactly as possible. In this paper we present a new application of a well-studied image coding technique, namely block truncation coding (BTC). It is shown that BTC can not only be used for color image compression, but can also be conveniently used for image recovery based image content databases. From the compressed stream BTC (without decoding), we derive two characteristics of image content description, one called the block color co-occurrence matrix (BCCM) and the other block pattern histogram (BPH).
A new application of a well-studied image coding technique, namely block truncation coding (BTC). It is shown that BTC can not only be used for color image compression, but can also be conveniently used for image recovery based image content databases. From the compressed stream BTC (without decoding), we derive two characteristics of image content description, one called the block color co-occurrence matrix (BCCM) and the other block pattern histogram (BPH). We use BCCM and BPH to calculate image similarity measures for content-based image retrieval applications. We present experimental results demonstrating that BCCM and BPH are comparable to similar prior art techniques.
To index color images, the features are extracted from the EDBTC which stands for error diffusion block truncation coding. EDBTC compression produces a bitmap image and two color quantizers, which are processed using vector quantization (VQ) to generate the image feature descriptor. Histogram characteristic (BHF) and color pattern histogram (CHF) function are calculated. The CHF is calculated from the color quantizer indexed by VQ and BHF is calculated from the bitmap image indexed by VQ. Based on these characteristics, similarity is measured between a query image and the database images. Finally, the image retrieval system returns a set of images to the user as output with a specific similarity criterion, such as texture and color similarities. Thus, the proposed EDBTC method is examined with good image compression capability and also offers an efficient way to index images for image retrieval.