Image Compression Technique for Low Bit Rate Transmission

2. Description Of The Proposed Technique

The process of the proposed ALIC technique can be described by the block diagram that is shown in Figure 1. At the transmitter, the input N×M image (N rows and M columns of pixels) is converted into a vector; the rows of the N×M matrix are concatenated sequentially in one long row that has NM length to form a single vector. This vector is subjected to DCT. A subsequent process is then applied to extract the highest energy coefficients of the DCT, thus leading to symbol reduction. DCTcoefficients that carry a specific energy percentage (EP%) of the image energy are selected to be quantized and the others are neglected. EP% is chosen so that the image can be restored from the selected coefficients without noticeable distortions in it. Then the reduced vector is subjected to one of the newly proposed quantization techniques. These quantization techniques work upon the histogram of the image, so they are adaptive techniques. The indices of the selected coefficients are stored to be used later.

The quantized coefficients are ordered according to their stored indices in a new vector that has the same length as the DCT input vector. This new vector is completed by zeros, and then are Run-Length encoded. The resultant vector is processed by Huffman encoder which generates coded output stream. In order to restore the original image again at the receiver, the previous stages are processed in the reverse direction. The following is the detailed description of the ALIC technique.