Energy thresholding based sub-band elimination DWT scheme for image compression

Images with texture patterns have embedded feature of offering visual patterns, that have the property of homogeneity. They provide cardinal information pertaining to the structural arrangement of the surfaces. Huge databases containing such images require highly competent compression schemes. This paper proposes a compression scheme for gray-scale texture images based on a unique Discrete Wavelet Transform (DWT) Sub-band Elimination guided by energy thresholding. The algorithm operates on sub-bands generated by the wavelet transform and determines the dominating coefficients which contribute to the maximum energy. It then eliminates redundant coefficients and facilitates efficient compression ratio. The performance of the algorithm is evaluated by calculating various quality metrics viz. PSNR, MSE etc and plotting apposite graphs by experimenting on standard texture gray-scale image database.