Template-Matching-based High-Speed Face Tracking Method using Depth Information

The present invention relates to a template matching-based high speed face tracking method which tracks a face from the next frame using a template after generating the template by detecting the face from a frame by receiving a depth image and a color image having the frame which is continuous in time from a camera photographing a human face. The present invention has a composition arranged to include the following steps of: (a) renewing the size of a template by comparing the depth of the template with the depth of a current frame; (b) setting the area of a current frame, expanded at a predetermined rate (expansion rate) of the size of the renewed template, as a search area around the renewed template; (c) selecting a matched position by performing a template matching in the search area using the renewed template; and (d) renewing the matched position to be at a position of the renewed template and repeating the (a) to (c) steps by setting the next frame of the current frame as a new current frame. The template matching-based high speed face tracking method as described above is able to improve the accuracy of a tracking by performing a matching correction while processing the tracking in real time by increasing the speed of the face tracking by reducing a search area by applying an early termination and a sparse search and adjusting the size of a template.