A principled approach to remove false alarms by modelling the context of a face detector

In this article we present a new method to enhance object detection by removing false alarms in a principled way with few parameters. The method models the output of an object classifier which we consider as the context. A hierarchical model is built using the detection distribution around a target sub-window to discriminate between false alarms and true detections. The specific case of face detection is chosen for this work as it is a mature field of research. We report results that are better than baseline methods on XM2VTS and MIT+CMU face databases. We significantly reduce the number of false acceptances while keeping the detection rate at approximately the same level.

• Type: Journal
• Year: 2010
• Where: BMVC
• Authors: Cosmin Atanasoaei, Chris McCool, Sébastien Marcel
• Keywords: BMVC 2010, Computer Vision, Detection, False Alarms, Object Classifier