An Improved Method of Image Compression with GPPM Tree and Haar Wavelet

Compression of image is done to reduce the size of the image, which reduces the size of the memory needed to store the image and the time required to transmit the image trough a network. There are two types of compression called the lossless compression and the lossy compression. Many works have been done in lossless compression. GPPM tree along with PPM tree are used in compression of RGB image in [1].This is a lossless compression. A small change is made in the steps involved in this existing algorithm. This step increases the compression ratio and decreases the bits per pixel. Haar wavelet is taken for the image and rounded values of the coefficients are taken and the algorithm in [1] is applied only on the LL component. For the other coefficients a mapping is done to convert the positive and negative coefficients into positive integer symbols. This is a one to one function. The Huffman coding is done for all the other three coefficients (LH, HL, and HH). Then all the compressed strings are combined into one file using a particular format. The testing is done with grayscale images. The original algorithm was implemented and checked for the performance of the compression algorithm in this paper. The metrics used for compression is compression ratio, PSNR, and SSIM and bits per pixel. Improved values are found for the compression ratio and bits per pixel.