Background Subtraction using gmm on single image

clc

clear all

close all

[file, pathname] = uigetfile('D:\studies\FYP\original images.jpg','Load Image');

cd(pathname);

foregroundDetector = vision.ForegroundDetector('NumGaussians', 3, ...

'NumTrainingFrames', 50);

face_original = imread(file);

size_img = size(face_original);

imshow(face_original), title ('Original Image')

foreground = step(foregroundDetector, face_original);

figure; imshow(foreground); title('Foreground');

this code gives black screen.

[Image Analyst]

I guess it didn't find a face. But what do I know -- you didn't post your image.

[Muzamil Hassan]

here is an image and this code gives output as an black image

Answer 1

[Aakash Mehta]

ForegroundDetector method is used to detect the motion in the video. As you are applying it to a single image, it is considering the whole image as a background.

foreground = step(foregroundDetector, face_original);

Above line will give the foreground mask which is 2D matrix. It contains 0's and 1's. In your case it's giving all 0's as it is not able to detect any foreground. That's why when you are trying to display image you are getting black image.

[Sarvath Sharma]

So how would one use the foregroundDetector and the step function to clear out and determine the 1s and 0s for the foreground mask? Would we have to do a comparison between two images and pass that in as a parameter to the step fucntion?