Automatic Face Detection and Tracking Using OpenCV Implementations of the Viola-Jones Method

Mihnea Horia Vrejoiu
mihnea@dossv1.ici.ro

Ştefan Alexandru Preda
stefanalex@ici.ro

Mădălina Zamfir
madalina@ici.ro
National Institute for Research & Development in Informatics, ICI Bucharest

Abstract: An experimental system has been designed, implemented and tested for human face detection and tracking in still, respectively dinamic images – succesive video frames from video files (movies) or captured in real time (live) with a video or even a web camera – based on the Intel’s open source computer vision functions library, OpenCV. In this context, there have been imagined, designed and implemented a scenario, a method and an algorithm for detecting and tracking each static or moving face thus identified in the respective frames, by tagging (indexing) it and computing its future possible positions. There have been provided solutions for correctly tagging / indexing the faces in successive frames including those possible situations of ambiguity in their correct identification in the case of temporarily „loosing” them from some frames and further reappearence.

Keywords: Face Detection, Face Tracking, Computer Vision, Image Analysis, Haar Cascade, OpenCV.

View full article