High Performance Hardware Architecture of Linear Filters for Intelligent Video Processing
PCM '08 Proceedings of the 9th Pacific Rim Conference on Multimedia: Advances in Multimedia Information Processing
| Hi-index | 0.00 |
In this paper, we develop a fast and efficient face image browsing system on CE (Consumer Electronics) devices. Our system adopts three methods such as facial region de- tection and facial feature extraction, facial vector cluster- ing, and DB handling for face metadata. Given photos, a facial region detection algorithm is applied to each photo and then Gabor-Wavelet features of fiducial points in each detected face are extracted. Next, a facial vector clustering algorithm makes all the facial features, which corresponds to faces, be clustered in an appropriate manner. In general, a face recognition algorithm requires a registration proce- dure, i.e., a user should register face images as the gallery, but our system clusters daily photos automatically. Finally, the DB is updated in order to browse photos by face meta- data in realtime. This paper presents the developed algo- rithms of the three methods in detail and an easy and effi- cient user interface for a face image browsing system on CE devices with its snapshots.