Image background correction for static object imaged using different background

Dear all,

Please I need help on background correction. I imaged a stationary object (person lying on a table) but with three different background scenes. The object position varies on each of the images despite the object was not moved (See image attached). The mean pixel value in the ROI (red rectangle) is 872mm, 843mm and 786mm for image a, b and c respectively.

Please is there any way I can correct the background so that the ROI value on all the images will be appproximately the same?.

Many thanks in anticipation of your help.

[Image Analyst]

Possibly. Upload the original image and original background image, not a screenshot of 3 pseudocolored images.

Answer 1

[Shreshth]

Hello Auwal,

It seems that[RS1] [SS2] you are dealing with background variation in your images, which can affect the mean pixel values in your region of interest (ROI).

You can solve the issue is by Background Subtraction process. It is a common technique used to separate foreground objects from the background in images. It works well when the background is relatively static. You can subtract the background image (without the object) from each of your three images to obtain a foreground mask. Then, compute the mean pixel value within your ROI on the masked images.

You can refer to this detailed example by MathWorks to understand the process of Background Subtraction:

https://www.mathworks.com/help/vision/ug/subtract-image-background-using-opencv-code-in-matlab.html

Hope it helps.