Performance evaluation of high quality image compression techniques

The area of image processing has offered remarkable ideas and competent compression algorithms in the past few decades. The accretion of data generated by applications validate the use of different image compression schemes to decrease the storage space and time for transferring the images over the link. Each of the compression schemes provide different image quality metrics viz. PSNR, MSE, Compression ratio, Normalized Cross Correlation, Normalized Absolute Error etc. This paper implements 3D-Discrete Cosine Transform (DCT) for compressing high resolution images with substantial amount of background. It presents the evaluation of the implemented algorithm, along with the performance comparison of other compression schemes viz. JPEG Lossy, Sub-Band replacement DWT and K-Means. It proves to be a dominant technique for multi-spectral images with improved quality parameters achieving 38-40dB PSNR.