Multiple classivvÖer object detection with convvÖdence measures.

This paper describes an extension to the Haar ClassivvÖer Cascade technique for object detection. Existing Haar ClassivvÖer Cascades are binary; the extension adds convvÖdence measurement. This convvÖdence measure was implemented and found to improve accuracy on two object detection problems: face detection and vvÖsh detection. For vvÖsh detection, the problem of selecting positive training-sample angle-ranges was also considered; results showed that large random variations that result in cascades covering overlapping ranges increases their accuracy.