2
$\begingroup$

Before today I had no idea there is any transform an such; I was only aware with DCT and wavelet transforms but today one of my teacher who is doing research on image processing project had told me that contourlet transform is way better than wavelet transform.

It is better because wavelet transform extract frequencies in 3 directions (diagonal, horizontal, vertical) while contourlet transform is capable of extracting information in any direction.

There are various books on wavelet transform but I couldn't be able to find a book that mainly focus on contourlet transform, is there any book you would recommend for contourlet transform?

P.S: I'm not talking about research papers.

$\endgroup$
  • $\begingroup$ What about PhD thesis of Minh N. Do? It's almost like a book. $\endgroup$ – DaBler Dec 21 '16 at 14:51
1
$\begingroup$

I do not know of any book solely devoted to contourlets. There is a chapter in Beyond Wavelets, 2003 (G. Welland, Ed.) by Do and Vetterli, named Contourlets. You also have a 100-page paper (almost a book) where contourlets are put in the perspective of Multidimensional Filter Banks and Multiscale Geometric Representations (2012).

At the risk of mildening your teacher's assertions, good compression is the result of a transformation with nice qualities and an efficient coding scheme. The "best" transform can fail to properly compress an image if you do not have a neat coding scheme associated with it. Contourlets, for instance, are slightly redundant. And its compaction capabilities are thus limited by the redundancy. To my knowledge, redundant transforms have not proved yet they are much better than critically-sampled ones for compression (except perhaps at large compression ratios). Even if you are not talking about research papers, let me mention a couple of different contourlet versions. The CRISP-contourlet, which is non-redundant, and The Nonsubsampled Contourlet Transform, fully redundant.

When trying to understand a transformation, especially with wavelet-like decompositions, it is often useful to look at the interplay between the math and the sampling schemes.

In addition to Minh N. Do thesis, I am trying to find a copy of A. L. da Cunha PhD.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.