Figure 3From: A multi-cue spatio-temporal framework for automatic frontal face clustering in video sequences Optical flow computed with pyramidal Lucas–Kanade. Back to article page