Figure 5From: A multi-cue spatio-temporal framework for automatic frontal face clustering in video sequences Preview of face detections, we use the OpenCV implementation of the Viola–Jones face detector. Face images are obtained by growing the detection to the under-head area.Back to article page