use the mouse or number keys 1–8 to switch between images

1: input


2: ground truth shadow-free


3: shadow mask


4: our output


5: Guo et al. 2012


6: Guo et al. 2012 (automatic detection)


7: Arbel and Hel-Or 2011