I have 2 types of images: those that are in the right orientation, and those that are not.
I want to detect if the image is in the right orientation based on the objects positions of it.
This is an example of a right image, and I highlited some of the objects that could be used to check the right orientation:
And here is the image in the wrong orientation:
Is there any way, using python, that could do this check to rotate the image to the right orientation based on the objects location of it?
I already tried using Structural similarity index from scikit image, but this methology doesn't worked, because there are a big differences of contrasts between the images , returning wrong results of the right orientation.