Summary: | ABSTRAK
Both wavelet packets (WP) and cosine packets (CP) have recently emerged as powerful tools for signal compression, particularly in areas of still image and video compression. This is a comparative study of performance on transform based coding image compression algorithms. The entropy-based criterion and optimal pruning algorithm are used to choose the best decomposition tree of CP and WP. To obtain a quantitative measure of our methods's performance, we give PSNR (peak signal to noise ratio), time-consumption and sum of zeros percentage parameters. The sum of zeros percentage parameter showed that WP will be the winner if the decomposition level is 5 and 6. Otherwise CP will be the winner if the decomposition level is 1 . This indicates WP is better than CP regarding to energy compaction. Other parameters proved that WP is better than CP.
Keywords: wavelet packets (WP), cosine packets (CP), PSNR, time-consumption, sum of zeros percentage.
|