how to retrieve pixel values from bounding box of an image.

Following is a code to create a bounding box of the detected face.

clear all

clc

faceDetector = vision.CascadeObjectDetector;

I = imread('visionteam.jpg'); % RGB IMage

figure

imshow(I)

title('Input Image')

bboxes = step(faceDetector, I);

IFaces = insertObjectAnnotation(I, 'Rectangle', bboxes, 'Face');

figure, imshow(IFaces), title('Detected faces');

--------------------------------------------------------------------

My question is:

How can I obtain pixel values inside this bounding box area?

Store the retrieved values in M by N by 3 matrix, after obtaining the pixel values under the bounding box area. How can it be done?

Answer 1

[Ameer Hamza]

See imcrop(): https://www.mathworks.com/help/images/ref/imcrop.html. For example, add following lines at end of your code

faces = cell(1, size(bboxes,1));
for i=1:numel(faces)
    faces{i} = imcrop(I, bboxes(i,:));
end

It will save each face in the cell array.

[sami ullah]

faces = cell(1, size(bboxes,1));

for i=1:numel(faces)

faces{i} = imcrop(I, bboxes(i,:));

end

After cropping, I have to change the each face pixel values with some random values and then place them (each face) on the original image bounding boxes and display the whole image with changed pixel values under boxing box.. How it can be done?.

[Ameer Hamza]

Try this code

faceDetector = vision.CascadeObjectDetector;
I = imread('visionteam.jpg'); % RGB IMage
figure
imshow(I)
title('Input Image')
bboxes = step(faceDetector, I);
IFaces = insertObjectAnnotation(I, 'Rectangle', bboxes, 'Face');
figure, imshow(IFaces), title('Detected faces');

% extract faces
faces = cell(1, size(bboxes,1));
for i=1:numel(faces)
    faces{i} = imcrop(I, bboxes(i,:));
end

% replace faces with random noise
for i=1:numel(faces)
    faces{i} = uint8(randi([0 255], size(faces{i})));
end

% put the random faces back
for i=1:numel(faces)
    c1 = bboxes(i, 1);
    c2 = sum(bboxes(i, [1 3]));
    r1 = bboxes(i, 2);
    r2 = sum(bboxes(i, [2 4]));
    I(r1:r2, c1:c2, :) = faces{i};
end
figure, imshow(I), title('Random faces');

[sami ullah]

Thanks a lot Ameer Hamza

[Ameer Hamza]

I am glad to be of help!