从（嘈杂的）纯色背景中提取单个对象

Question

I'm in the process of scanning old photographs, and I would like to automate the process of extracting the photograph from the (noisy) solid white background of the scanner so that I have a transparent photograph.

I've tried to use OpenCV in Python, with techniques as edge/contour detection, but it would pick up the little specs of dust in the background and not generate a complete and correct transparency map.

I have a sample:

and the desired result:

Answer 1

there are the steps necessary to achieve your desired result (I implement them in C++ so I don't attach the code):

1. Convert your image to grayscale.
2. Binarize it with a reasonably high threshold (I use 240).
3. Optionally apply opening operation.
4. Find contours.
5. Get the biggest contour.
6. Draw this biggest contour on another Mat filled with zeros
7. You get the transparency mask