Two Phased Decompression Scheme for Quality Improvement of Vector Quantized Images

With the development of the cloud virtual environment, a lot of storage space to store the user´s image files is looking forward in the future. in this study, we hope to use the user already owned image database to enhance the quality of a VQ decompressed image which is received its index table by the user from internet. the major idea is to set an indexing structure for all of the images in database. the indexing structure is established according to the VQ codebook such that image blocks (in size 4x4 as usual) will be linked with the same codeword if this codeword is their most similar codeword in the VQ codebook. All the blocks were clustered into groups in accordance with the same index value of codeword, any compressed block using the same codeword will correspond to a clustered group to find more suitable block to restore this image block. the experimental results showed that our method can restore the decompressed images in better quality from the same index table as the conventional VQ decompression.