How can the image Processing of .tif file of size 763x1023x4 be done?

I am using SEM files in .tif format for image processing using MATLAB to extract the edge, size, and area of the microstrtucture. The size of the file is 763x1024x4, The image is already in grayscale colour and I am not able to apply the command which I am leaning through Matlab online course. there are many errors showing just using imshow(file) command like

Error using images.internal.imageDisplayValidateParams>validateCData

Multi-plane image inputs must be RGB images of size MxNx3.

Error in images.internal.imageDisplayValidateParams (line 30)

common_args.CData = validateCData(common_args.CData,image_type);

Error in images.internal.imageDisplayParseInputs (line 79)

common_args = images.internal.imageDisplayValidateParams(common_args);

Error in imshow (line 253)

images.internal.imageDisplayParseInputs({'Parent','Border','Reduce'},preparsed_varargin{:});

Error in sample (line 5)

imshow(I)

Kindly help me with solution, since I am stuck at beginning, so could not proceed with image segmentation, detection & analysing texture.

[Image Analyst]

What is the format of the file -- what is it's filename extension?

Have you looked for readers for that file type in the File Exchange?

Can you attach the file here?

How did you read in the file to (the badly-named) I? Or is (the badly-named) I a character array that is the filename rather than the image variable itself?

Answer 1

[DGM]

Imshow() can only represent 1-channel (grayscale) images or 3-channel images (represented as RGB).

So whatever you do to display a 4-channel image will depend on what the channels represent. Normally, if I see a 4-channel image, I expect that it's RGBA or CMYK. Alternatively, it could be some sort of multispectral image, so as with RGBA you'd have to know which channels are which in order to know how to split/view it. Maybe it's just a volumetric image with only 4 z-slices. Maybe it's not even conceptually volumetric, but merely four independent grayscale frames.

I'm not sure what conventions exist for SEM imaging. I'm guessing it's four independent grayscale frames. If that's the case, then you can just show the pages independently:

imshow(A(:,:,1)) % show just the first page/frame

Alternatively, you can use montage for a quick view of all pages.

A = imread('cameraman.tif');
A = cat(3,A,rot90(A,1),rot90(A,2),rot90(A,3)); % make a 4-channel image to test

% swap dim 4 and 3 so that montage correctly treats it as a multiframe image
% montage arranges the images row-wise (left-right, top-bottom)
montage(permute(A,[1 2 4 3]))

[NITIN KUMAR]

@DGM Thank you very much for the solution, now I am able to view the picture in .tif format but still have to change it from 4 to 3 so that image segmentation can be performed. also I am not able to open it in image viewer or image segmentation app. Could you advice?

[DGM]

Trying to reduce it to 3 channels presumes that it's RGBA. Is it RGBA? If it's not, then it wouldn't make any sense to reduce it to RGB, since the first three channels are not visible color information. If it's a stack of grayscale frames, then you'd treat it as four grayscale images and do segmentation on each one. I don't know what your images conceptually represent. That's something you'll have to figure out. Either way, splitting the image is simply a matter of array indexing like in the first example.

B = A(:,:,1); % get the first page
C = A(:,:,2); % get the second page
D = A(:,:,1:3); % get the first three pages
% etc

Since preprocessing is going to be required, you'll have to import the image into the workspace, process it as needed. At that point, you should be able to get the segmentation app get the prepared image(s) from the workspace instead of trying to read it from the file directly.

[NITIN KUMAR]

@DGM Thank you so much for the help, I will try out it now, I am facing a problem regarding the digitalization of micrographs so that those can be used to generate approx geometry of the microstructure of the ceramic material to perform simulations for obtaining knowledge about fracture toughness. I want to create a geometric/model file in the format of any cad software so that it can be further converted to be used in Abaqus software.

The micrographs are in .tif format and do not contain vector information for processing into Abaqus.

[DGM]

The description of the task of extracting geometry information from a raster image and converting it to some vectorized representation is still a bit too abstract to approach with a concrete example.

The vectorization/triangulation/export part isn't something I'm really experienced with, and I'm not sure how much flexibility is afforded by the importing/preprocessing capabilities of Abaqus. I'm definitely not the best person to ask about that.

I will say that for the exporting part itself, you may be able to use stlwrite() or igesout() (File Exchange)

https://www.mathworks.com/matlabcentral/fileexchange/14470-igesout

[NITIN KUMAR]

@DGM thank you so much for zour advice, I have started working on part by part on the image to generate approx vector details so that i can have an approy geometry. I will also check out the link which you sent.

Answer 2

[Image Analyst]

If you have a hyperspectral image and want to know how to process and display as an RGB image, see the attached article.