Resizing a binary image issue

Question

Aaron Devanathan 2021-5-5

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/822370-resizing-a-binary-image-issue

编辑： Image Analyst 2021-5-5

BW_05052021.png

Hello. I am trying to resize a binary image to match the dimensions of another workspace matrix. Yet I am losing the non-zero values in the new, resized matrix. I have attached the binary image ("BW_05052021.png") that serves as a mask.

%The value of the BW image is 10000 x 8082 logical
%Looking to resize the image matrix to 142 x 128, not worried about maintaining scalar ratio
nnz(BW); % ans = 952
BW_resize = imresize(BW, [142 128]); % Resizing BW to the specified dimensions; 142 rows and 128 columns
nnz(BW_resize); % ans = 0 and not sure why

If needed, I can clarify or provide additional information as needed. Thank you!

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Atsushi Ueno 2021-5-5

It looks like the original image is lost because the ratio of resizing is too high. If you compare the original image with images that have been reduced to 75%, 50%, and 25%, you can see the limit of the reduction ratio at which the original image can be recognized. You may need gray-scale image for your requirement.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Walter Roberson 2021-5-5

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/822370-resizing-a-binary-image-issue#answer_692825

编辑：Walter Roberson 2021-5-5

在 MATLAB Online 中打开

This is to be expected.

You are compressing by a factor of 70 for rows and 63 for columns, so each output pixel is calculated by merging information from over 4400 pixels. If you were to take the mean over the area and say that it is to be set if the mean is > 1/2 then you would be far far far short for any pixel.

BW = imread('https://www.mathworks.com/matlabcentral/answers/uploaded_files/608410/BW_05052021.png');

imshow(BW)

rb = floor(size(BW,1)/142);

cb = floor(size(BW,2)/128);

reduced = cast(blockproc(BW, [rb, cb], @(block) max(block.data(:))), class(BW));

size(reduced)

ans = 1×2

143 129

BW_resize = imresize(reduced, [142 128]);

imshow(BW_resize)