I would like to ask the logic behind the shape distance provided in opencv 3.1 as I am confused by the results of using shape distance when comparing duplicated images.
I used the shape distance example as the starting point and tested the shape distance between copied bmp images (I just duplicated the bmp file of the sample image). The shape distance calculated is not stable and not equal to zero which makes me confused of the meaning of the magnitude of the calculated shape distance.
Therefore, I would like to ask for the logic behind the shape distance calculation. Thanks!