everybody knows that in transform coding the image to be coded is at first converted in the transform domain.this transformation has two effects.first,compression and second,decorrelation. can anybody explain to me very precisely how this decorrelation helps anyway (preferably, from 'image coding' point of view)? |
|
transform coding...decorrealtion...
Started by ●September 23, 2002
Reply by ●September 24, 20022002-09-24
Decorrelation is analytically the same operation as the KLT. In fact KLT can
be explained or derived using two approaches: 1. decorrelation, 2. energy compaction. If you start deriving KLT with the motivation of obtaining a transform which produces K out of M samples with most energy stored in the K samples, and if you do this for all K from 1 to M, you come up with the optimal transform; and at the same time you see that this transform is the one that decorrelates the data. The above derivation of KLT is more likely to be found in pattern recognition texts where they try to obtain the smallest representation of the data (the feature vector) with the most amount of information. So, if the intuitive link between decorrelation and compression (which is directly related to compaction in representation) is not clear for you, I can recommend the above analytical point of view. Omer ----- Original Message ----- From: k_2002_m To: Sent: Monday, September 23, 2002 3:20 PM Subject: [imagedsp] transform coding...decorrealtion... everybody knows that in transform coding the image to be coded is at first converted in the transform domain.this transformation has two effects.first,compression and second,decorrelation. can anybody explain to me very precisely how this decorrelation helps anyway (preferably, from 'image coding' point of view)? _____________________________________ Note: If you do a simple "reply" with your email client, only the author of this message will receive your answer. You need to do a "reply all" if you want your answer to be distributed to the entire group. _____________________________________ About this discussion group: To Join: To Post: To Leave: Archives: http://www.yahoogroups.com/group/imagedsp More Groups: http://www.dsprelated.com |
Reply by ●October 11, 20022002-10-11
decorrelation means that the correlation between the valuesin the transformed domain is reduced or may be made zero. this i think will reduce the no. of bits required to code the image i.e there is no redundancy in the details and so fewer coefficients arise. bye nitthilan --- k_2002_m <> wrote: > everybody knows that in transform coding the image > to be coded is at > first converted in the transform domain.this > transformation has two > effects.first,compression and second,decorrelation. > > can anybody explain to me very precisely how this > decorrelation helps > anyway (preferably, from 'image coding' point of > view)? > |