Range Face Segmentation: Face Detection and Segmentation for Authenticationin Mobile Device Range Images

Abstract

Face detection (finding faces of different perspectives in images)is an important task as prerequisite to face recognition. This isespecially difficult in the mobile domain, as bad image quality andillumination conditions lead to overall reduced face detection rates.Background information still present in segmented faces and unequallynormalized faces further decrease face recognition rates. We presenta novel approach to robust single upright face detection and segmentationfrom different perspectives based on range information (pixel valuescorresponding to the camera-object distance). We use range templatematching for finding the face’s coarse position andgradient vector flow (GVF) snakes for precisely segmenting faces.We further evaluate our approach on range faces from the u’smileface database, then perform face recognition using the segmentedfaces to evaluate and compare our approach with previous research.Results indicate that range template matching might be a good approachto finding a single face; in our tests we achieved an error freedetection rate and average recognition rates above 98%/96% forcolor/range images.

Publication
Proc. MoMM 2013: 11th International Conference on Advances in MobileComputing and Multimedia