Image Compression using Self Organizing Map and Discrete Wavelet Transform with Error Correction

The compression is a vital field in image processing which reduces the bandwidth during data transfer in multimedia communications. In this paper, we propose Image Compression using Self Organizing Map (SOM) and Discrete Wavelet Transform (DWT) with Error Correction. The input image is considered from standard database and modified to 4x4 blocks and rearranged as 16x4096 input vectors and applied to the input layer of SOM. The SOM produces index and weight vectors where index vectors are repeated values of the class and weight vectors are the trained values. The Index vectors are then compressed to obtain index bits using arithmetic encoder and the weight vectors are applied to DWT to obtain first level compression of LL band weight vectors and further encoded using quantizer and arithmetic encoder to obtain second level of compressed weight bits. The Error Matrix is obtained and is further Quantized and arithmetic encoded to obtain compressed Error bits. The compressed weight and index bits along with compressed error bits are transferred over the channel. Finally, the original image is reconstructed at the receiver using decompression algorithm. The performance of the algorithm is tested with standard set of images and compared with existing techniques to validate the results. It is observed that the proposed method is better compared to existing techniques in terms of PSNR.