For best tree decomposition using wavelet packets, i have tried both Shannon and log energy entropy functions. I noticed that log energy leads to images with higher PSNR values. Could you please help me explaining the reason?

  • The Matlab expression for the two cases is as following: [Arec,bitstream,PSNR,MSE,D,Drec,s,p_stream]=spiht_wpackets('lena256.bmp',1,'CDF_9x7',5,4,'greedy','shannon'); [Arec,bitstream,PSNR,MSE,D,Drec,s,p_stream]=spiht_wpackets('lena256.bmp',1,'CDF_9x7',5,4,'greedy','logenergy');

Lena256 (image with size 256*256)

  • $\begingroup$ do you have example pictures? $\endgroup$ – Marcus Müller Mar 6 '18 at 17:56
  • $\begingroup$ For Lena image, Shannon entropy results in 33.3474 PSNR, and 37.0216 PSNR for logarithm energy entropy. $\endgroup$ – user24907 Mar 6 '18 at 18:17
  • $\begingroup$ wait. Are we talking about the same image, just two different PSNR definitions, or are we talking about two different reduced versions of the same image, being subjected to the same PSNR measurement method? $\endgroup$ – Marcus Müller Mar 6 '18 at 18:54
  • $\begingroup$ @ Marcus Müller For the same image, the same PSNR measurement method and the same bpp: the use of Shannon entropy achieved 33.3474 PSNR, and log energy achieved PSNR of 37.0216. $\endgroup$ – user24907 Mar 6 '18 at 19:31
  • $\begingroup$ what is "use of shannon entropy", then? I'm confused. Can you add a description of what exactly you're doing to get these two values to your question, please? $\endgroup$ – Marcus Müller Mar 6 '18 at 20:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.