You can get the point using the same transformation matrix that applies to the image. You want to convert the point between the screen coordinate system to the image coordinate system by changing the effect of the original matrix.
In particular, you want to convert the x, y coordinates, where the user clicked on the screen at the corresponding point in the original image, using the inverse matrix that was used to convert the image to the screen.
A bit of pseudo-code assuming a matrix contains a transform that was applied to the image:
source share