Automatic Face Location to Enhance Videophone Picture Quality

New video communication and multi-media products open up a range of machine vision applications, in which the potential size of the market can justify a substantial investment in the development of sophisticated algorithms. Face location can be used to enhance the subjective performance of videophones, while still conforming with international video compression standards. This paper gives an overview of the location techniques employed, describes a real-time implementation, and presents the results of the subjective tests which conftrmed the improvement in picture quality.