“…The methods are classified in groups based on their model structures, and the performance of the different model-groups are compared via test sequences. The authors note that the methods work in different color spaces, like RGB (Mikic et al, 2000) and HSV (Cucchiara et al, 2001), however, it remains open-ended, how important is the appropriate color space selection, and which color space is the most effective regarding shadow detection. Moreover, we find also further examples: (Rittscher et al, 2000) used only gray levels for shadow segmentation, other approaches were dealing with the CIE L*u*v* (Martel-Brisson and Zaccarin, 2005), respectively CIE L*a*b* (Rautiainen et al, 2001) spaces.…”