A novel video face clustering algorithm based on divide and conquer strategy

Authors:
Gaopeng Gou;Di Huang;Yunhong Wang
Affiliations:
Laboratory of Intelligent Recognition and Image Processing, Beijing Key Laboratory of Digital Media, Beihang University, Beijing, China;Laboratory of Intelligent Recognition and Image Processing, Beijing Key Laboratory of Digital Media, Beihang University, Beijing, China;Laboratory of Intelligent Recognition and Image Processing, Beijing Key Laboratory of Digital Media, Beihang University, Beijing, China
Venue:
PRICAI'12 Proceedings of the 12th Pacific Rim international conference on Trends in Artificial Intelligence
Year:
2012

Citing 2
Cited 0

Distinctive Image Features from Scale-Invariant Keypoints

International Journal of Computer Vision
Content based indexing of images and video using face detection and recognition methods

ICASSP '01 Proceedings of the Acoustics, Speech, and Signal Processing, 2001. on IEEE International Conference - Volume 03

Quantified Score

Hi-index	0.00

Visualization

Abstract

Video-based face clustering is a very important issue in face analysis. In this paper, a framework for video-based face clustering is proposed. The framework contains two steps. First, faces are detected from videos and divided into subgroups using temporal continuity information and SIFT (Scale Invariant Feature Transform) features. Second, the typical samples are selected from these subgroups in order to remove some non-typical faces. A similarity matrix is then constructed using these typical samples in the subgroups. The similarity matrix is further processed by our method to generate the face clustering results of that video. Our algorithm is validated using the SPEVI datasets. The experiments demonstrated promising results from our algorithm.