imcrop not working for tiff

Hi, i want to crop several .tiff pics so i wrote below code but it dosnt work for .tiff files. it works for .png .jepg . i used imshow it shows nothing for .tiff files.

clc;
clear all;
address='D:\all\uni things\1-master\master project\6-Data\1-pics\1-main\2\L\2-E\5';
filenames=dir(fullfile(address,'*.tiff'));

for n=1:numel(filenames)
fullname=fullfile(address,filenames(n).name);
t1=imread(fullname);
info = imfinfo(fullname);
t2=imcrop(t1,[10, 10, 1000, 950]);
t3=size(t2);
imwrite(t2,fullname);
end

[Mann Baidi]

Hi,

Please share the files and the error which you are facing

[Umar]

Hi Ali,

I have made modifications to your code. Below is the updated code with added cropping and display functionality:

% Define the directory containing the .tiff images
address = 'D:\all\uni things\1-master\master project\6-Data\1-pics\1-main\2\L\2-E\5';

% Get a list of .tiff files in the directory
filenames = dir(fullfile(address, '*.tiff'));

% Loop through each image file
for i = 1:length(filenames)
    % Read the image
    t = Tiff(fullfile(address, filenames(i).name), 'r');
    imageData = read(t);
    
    % Perform cropping operations on the image
    % Add your cropping code here
    % For example, crop the image to a specific region
    croppedImage = imcrop(imageData, [x y width height]);
    
    % Display the cropped image
    figure;
    imshow(croppedImage);
    title('Cropped Image');
    
    % Save the cropped image (optional)
    imwrite(croppedImage, 'cropped_image.tif');
end

For more information on Tiff and read functions, please refer to

https://www.mathworks.com/help/matlab/ref/tiff.html?searchHighlight=Tiff&s_tid=srchtitle_support_results_1_Tiff

https://www.mathworks.com/help/matlab/ref/tiff.read.html?s_tid=doc_ta

I hope this answers your question.

[DGM]

It's hard to be sure what the problem was, but there are a couple other possibilities to consider.

A TIFF might be indexed color. Without reading the color table from the file, the index array may not render in a manner which appears correct or is easily visible.

% unzip the sample tiffs
unzip tiffbucket.zip

% an indexed image with a short colortable (12 colors)
[inpict,CT] = imread('indexedblobs.tiff');

% displaying the index array by itself appears black
% because the dynamic range is only [0 11] versus [0 255]
% as would be expected of a uint8 intensity image
imshow(inpict)

% but displaying the image with its color table fixes that
imshow(inpict,CT)

% similarly, if the goal is to save a copy, we need the color table
imwrite(inpict,CT,'myblobs.tiff')

Similarly, a TIFF might not contain full-scale image data. It's not uncommon for raw images to be represented in wider integer formats than the actual dynamic range of a specific image may require. Avoiding preprocessing is part of the appeal of capturing raw data, after all. If your image is (e.g.) uint16, but the data itself spans a smaller interval (e.g. [1000 2000], then the raw data will be rendered on screen as a nearly-black image.

% arbitrary-scale data stored in uint16
inpict = imread('cman_arbrange_u16.tiff');

% the dynamic range of the data isn't tied to the numeric class
imrange(inpict)

ans = 1x2

1000 2000

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

% scale-dependent tools expect black and white to be implied by the class
% in this case, black is at 0, and white is at 65535
% so the image is almost black (hard to see against the white background)
imshow(inpict) % use native-scale (uint16-scale)

% for viewing, you can use the DisplayRange parameter
% (this only works for single-channel images)
imshow(inpict,[]) % scale to data extrema

% or you can normalize the image using mat2gray(), rescale(), etc
outpict = mat2gray(inpict); % normalize to unit-scale float
outpict = im2uint16(outpict); % maybe we want to keep the original class

% at that point, the image could be saved
imwrite(outpict,'cman_rescaled.tiff')

Again, it's hard to know what the problems might have been, but these are fairly common pitfalls. If you run into images which unexpectedly appear black (or white), it's a good start to check that the data scale is appropriate for its class. How that should be resolved depends on the goals and order of operations .

Answer 1

[Karan Singh]

Hi @Ali yousefi, I think @DGM has solved the query, for the crop part I dont think there is any mistake, here I have tried to keep a check for the same and the images are not empty so you can move forward with your origional code.

clc;
clear all;
unzip tiffbucket.zip
    
[t1,CT] = imread("indexedblobs.tiff");

if isempty(t1)
    disp(['Failed to read image: ', fullname]);
end

figure;
imshow(t1,CT);
title(['Original Image: ', "indexedblobs.tiff"]);

t2 = imcrop(t1, [200, 10, 2000, 950]);
figure;
imshow(t2,CT)