A new criteria for evaluation of compression technique

Currently, most of the parameters used to evaluate the performance of a compression technique are statistical based like, average error, RMS error, etc., which provide the impact of image compression on a global basis, but does not indicate the effect on edges in any region of the image. No direct single parameter is available which gives the quantitative measure of the effect of compression on the strength of edges in a selected region. Here, a new parameter called the frequency capture function (FCF) has been proposed to measure the performance of a compression technique in terms of the impact on the edge strength in a selected region of the image. We have used this parameter to compare the performances of JPEG, wavelet based set partitioning in hierarchical trees (SPIHT) and adaptive differential pulse code modulation (ADPCM) compression techniques for a test image. Results obtained are in agreement with visual quality evaluation of the images compressed and reconstructed with these techniques.