This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
Face detection demo is based on the Viola & Jones approach. This approach takes a sliding window approach over the input and tests if the window can be classified as a face. If so, the window is stored as a bounding box of a detection.
As a cleaning step, the algorithm goes over all bounding boxes and sees where there is a lot of overlap. There he creates a single bounding box with average dimensions.
So the end result is a set of bounding boxes, containing coordinate of upper left corner, the width and the height, according to the openCV coordinate system.
