How to apply a quantization step to a 2D transformed image?

4 次查看（过去 30 天）
Said BOUREZG
Said BOUREZG 2016-2-5
评论： Walter Roberson 2016-2-6
Hi all of you, I need some help here, my problem is how can i use quantization to the result of that code below?,knowing that the result (XDFT) is complex arry, example: XDFT(1,1)=125+ i*2562. So how values like XDFT(1,1) will be on an interval between 0 and 255.
[filename, pathname] = uigetfile('*.bmp', 'open image');% getting a Bitmap image.
[X,MAP]=imread(fullfile(pathname, filename));%Image reading.
R=input('R=');%bloc size(8*8 or 16*16 or 32*32).
XDFT= blkproc(X,[R R],'fft2');%DFT Transformation of X.

请先登录，再回答此问题。

采纳的回答

Matt J
Matt J 2016-2-5
Why not quantize the real and imaginary parts? Or, like in JPEG compression, why not use the DCT, instead of the DFT, so that the block spectra are real-valued?
  2 个评论
Said BOUREZG
Said BOUREZG 2016-2-6
I don't have a problem with DCT or DWT, my problem with DFT coefficients if I separate real and imaginary parts and I quantize them, the compression of them can't give a good results.
Walter Roberson
Walter Roberson 2016-2-6
Why is it not possible for the compression to give good results? If you keep 5 real components and 5 imaginary components, that is as good a compression as keeping 10 real components.

请先登录，再进行评论。

更多回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Image Segmentation and Analysis 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by