How to convert video to images (0~255) using background subtraction.

Hello.

I wanna convert video to images (0~255).

When I used this code, I just got 0 or 1 (logical).

Can I get 0~255 instead of 0 or 1 (logical)? Thank you!

clear all
close all

%// read the video:
reader = VideoReader('lena_side.avi');
vid = {};
while hasFrame(reader)
    vid{end+1} = im2single(readFrame(reader));
end
%// simple background estimation using mean:
bg = mean( cat(4, vid{:}), 4 );
%// estimate foreground as deviation from estimated background:

for i=1:60
    fIdx(i) = i; %// do it for frame 1 ~ 60
    fg{i} = sum( abs( vid{fIdx(i)} - bg ), 3 ) > 0.25;
    fg{i} = reshape(fg{i},[],1);
end
D1 = cell2mat(fg);

Answer 1

[Ameer Hamza]

Use im2uint8: https://www.mathworks.com/help/releases/R2019b/images/ref/im2uint8.html

imgUINT8 = im2uint8(img0_1);

[Kong]

Thank you so much!

I got this error. Do you have an idea?

clear all
close all
%// read the video:
reader = VideoReader('lena_side.avi');
vid = {};
while hasFrame(reader)
    vid{end+1} = im2uint8(readFrame(reader));
end
%// simple background estimation using mean:
bg = mean( cat(4, vid{:}), 4 );
%// estimate foreground as deviation from estimated background:
for i=1:60
    fIdx(i) = i; %// do it for frame 1 ~ 60
    fg{i} = sum( abs( vid{fIdx(i)} - bg ), 3 ) > 0.25;
    fg{i} = reshape(fg{i},[],1);
end
D1 = cell2mat(fg)

[Ameer Hamza]

The error is caused by difference in types of vid{fIdx(i)} (uint8) and bg (double). Change the calculation of bg with following line

bg = uint8(mean( cat(4, vid{:}), 4));

[Kong]

Thank you so much!

Do you know the meaning (> 0.25;) ? what is the 0.25?

fg{i} = sum( abs( vid{fIdx(i)} - bg ), 3 ) > 0.25;

[Ameer Hamza]

I am not sure about it, but it seems like a threshold, below which the values are filtered. By inspecting the matrices, i found that the term sum( abs( vid{fIdx(i)} - bg ), 3 ) is uint8, so its value are in range [0-255]. Therefore comparison with 0.25 does not make much sense. I think 0.25 was meant when pixel intensity is scaled between [0-1]. So it seems that you should multiply it by 255, i.e., use > 0.25*255.