0 个投票

I want to divided image to non-overlapping blocks, then I will appliy DWT on each block
can any one help me with this please

请先登录，再回答此问题。

 采纳的回答

Image Analyst
Image Analyst 2019-3-9

0 个投票

This is a FAQ, so See the FAQ
I recommend blockproc as the best for what you want to do. I'm attaching demos for non-overlapping blocks.

4 个评论
显示 2更早的评论 隐藏 2更早的评论

Mohammed Kassem
Mohammed Kassem 2019-3-21
how to perform DWT on each block of image using (blockproc)
i dont need the standard deviation can you help me please
thanks.
Image Analyst
Image Analyst 2019-3-21
Use the demo where I write my own custom function to process each block. You can do whatever you want in that function, including DWT or anything else.
Mohammed Kassem
Mohammed Kassem 2019-3-23
Thanks for your efforts i have changed the function as
I = imread('pears.png');
fun = @(block_struct)...
dwt2(block_struct.data,'dmey');
C = blockproc(I,[64 64],fun);
but i got no out put and no errors
so how i can access the [cA,cH,cV,cD] of dwt2 if the above code is true
Image Analyst
Image Analyst 2019-3-24
I can't help you with the wavelet part since I don't have the wavelet toolbox.

请先登录，再进行评论。

更多回答（1 个）

KALYAN ACHARJYA
KALYAN ACHARJYA 2019-3-9
编辑：madhan ravi 2019-3-9

1 个投票

You can use blockproc (Distinct block processing for image) on image
result=blockproc(image,[m n],fun)
Which processes the image by applying the function fun to each distinct block of size [m n] and concatenating the results into the output result matrix. You have to create the function
See the documentation for detail
Also you can look for sliding window (one by one pixel), check
kernel=create window
result=conv2(image, kernel, 'same');
If it still unclear, let me know?

类别

在 帮助中心 和 File Exchange 中查找有关 Neighborhood and Block Processing 的更多信息

产品

标签

Community Treasure Hunt

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

Start Hunting!

Translated by