Extracting background/foreground of an image

Hey Guys, I am currently starting to get into matlab and want to experiment with some images. So basiccaly I load two images (the same, but different angle) and I want to extract their background or foreground. So basically if I detect a human in the foreground, it should „extract“ him. (The goal in the end, is to place the extracted foreground on a different background) But how do I find this human in the background. Should I do some correspondence calculation with harris detector or something like that? Should I transform the images into grayscale? What is the best way to calculate this mask to extract the features of an foreground/background. Is there a way without using some toolboxes? If not, what kind of toolboxes would help. Thank you very much Guys !!

[Walter Roberson]

https://www.mathworks.com/matlabcentral/fileexchange/49503-moving-object-tracking-based-on-background-subtraction

Answer 1

[Image Analyst]

If your background is static, then you can get one with no humans and subtract it from the test image and look for differences.

[Tony Rochelle]

Yes that is the easiest solution of course thank you. But I want to loop through multiple images and some of them doesnt have humans in it. How do I then apply the subtraction? It is like a video which is separated into the frame images

[Image Analyst]

Do it the same way. Just subtract the background. It doesn't matter if there are people in it or not. Why do you ask? Do you think it does matter? If so, why?

Attached is a more complicated version that uses a video and computes a recent background by using a decreasingly weighted sum of prior frames. Adapt as needed.

[Tony Rochelle]

Okay thank you, when subtracting the image, how can I get the mask. Is the mask the difference of these images? And how do I place this extracted foreground into another background?

[Image Analyst]

I believe I answered that in another question of yours.