How can I select the gray patches

I have this picture and I tried to wrote a code to select these 12 grey patches. The problem I am facing is that some patches are not detected properly. How can I correctly select these 12 ROI.

Fig 1: Input Image Fig 2: It miss the 2 ROIs and have false detection.

CropImg=imread('patchc.jpg');

img=rgb2gray(CropImg);

rangefiltv= rangefilt(img);
imshow(rangefiltv)
mask = rangefiltv >10;
mask=~mask;

mask = bwareafilt(mask, 12);
stat = regionprops(mask,'boundingbox');

imshow(CropImg); hold on;
for cnt = 1 : numel(stat)
    bb = stat(cnt).BoundingBox;
    rectangle('position',bb,'edgecolor','r','linewidth',2);
end

*Update

The Cropped image is from this image. The size of the picture may varry.

Answer 1

[Image Analyst]

This looks like a computer graphics image where the gray patches will be in exactly the same location for any similar image, so you can just hard code in the coordinates of the corners. If that's not true, then post an actual image, not a perfect computer graphics one, so we can write segmentation code for it. Something like

mask = grayImage ~= backgroundGrayLevel;

[Masood Salik]

I have updated the full picture. The gray patches will be in exactly the same location for any similar image but the size of image may varry. Also the boarder may not be this clean. So I have cropped it.

If I fix the cordinates then how can it vary with varing size of picture?

[Masood Salik]

One of the patch colour is same as bacground, so if I use

mask = grayImage ~= backgroundGrayLevel;

Then that particular patch get removed. I think hard code coorinates method can work, but I don't know how to do that with varing image size.

[Masood Salik]

I have seen some MATLAB GUI that do such type of work that I wanted to do. How can I achive this ?