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