how to remove four black frame of image

Hi All,

I have an image and I want to remove four black frame of its corners without to remain any black pixel(i.e skin only ) in the image,please could any one explain to me using Matlab code how to do that without cropping ,using masking? thanks in advance.

[hamed abdulaziz]

[hamed abdulaziz]

the image after resizing

Answer 1

[Image Analyst]

If you take an image of a white scene, then you can divide your image by that to reduced or eliminate the vignetting.

[hamed abdulaziz]

Image Analyst: nice new icon,I downloaded this image and others from a data bast from medical site on the internet,and I want to segment the lesion in it for diagnosis,so that could help again, thank you in advance.

[hamed abdulaziz]

Image Analyst,I am waiting your answer

[Image Analyst]

Can you just take the blue channel and threshold it? Or, use the delta E app in my File Exchange. Or run a texture filter on it like stdfilt() or entropyfilt() and then threshold. Or else look for published algorithms here: http://iris.usc.edu/Vision-Notes/bibliography/contents.html

[hamed abdulaziz]

I wrote this piece of code:

Image1 = imread('d:\10.bmp');

blue=Image1(:,:,3);

level=graythresh(blue);

bw=im2bw(blue,level);

figure, imshow(bw);

but I get this image

[hamed abdulaziz]

[hamed abdulaziz]

then I tried

Image1 = imread('d:\10.bmp');

j=entropyfilt(Image1);

level=graythresh(j);

bw=im2bw(j,level);

figure, imshow(bw);

but no reasonable result!

[hamed abdulaziz]

Image Analyst,I am waiting your pretty answers for difficult questions as we are accustomed.

[hamed abdulaziz]

Why no answer?

[Image Analyst]

I'm at work, doing my real job. I'm not a paid Mathworker who helps people in this forum all day. And I'm sure it's going to be a lot more complicated than just some 3-line piece of code.

[hamed abdulaziz]

I am sorry to bother you ,I wish to you more and more of success in your job because you are smart person,and if you have a moment of the time,I am waiting your answer. Best regards

[Image Analyst]

Try inverting the binary image, then do a hole fill, then extract the biggest blob (see my code for that attached). You might also try activecontour() - a demo for activecontour is also attached.

[hamed abdulaziz]

Thank you very much,I tried inverting the binary image, then did a hole fill, then extract the biggest blob using activecontour_demo ,I got the below result but no lesion only please guide me the modified activecontour_demo is also attached for reviewing if you have a time?

[Image Analyst]

See my attached code.

[hamed abdulaziz]

Image Analyst : Thank you very much it is so useful code and excellent work.