A New Method for Multi-view Face Clustering in Video Sequence

Huang, Panpan; Wang, Yunlong; Shao, Ming

doi:10.1109/icdmw.2008.63

Cited by 9 publications

(7 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Antonopoulos, et al [8] proposed using a dissimilarity matrix from pre-existing face clusters; Fizsgibbon and Zisserman [9] used joint manifold distance; Tao and Tan proposed dividing face sequences by pose and then applying additional constraints obtained from domain knowledge [10]. Other methods that have been proposed in [11][12] [13] used mutual information, generative models with SVMs, and cluster consistency in videos to improve recognition performance. Finally, given the challenges of applying face recognition to personal photo collections, O'Hare and Smeaton [15] proposed the use of context information to more accurately identify people.…”

Section: Related Workmentioning

confidence: 99%

VidCat: an image and video analysis service for personal media management

Begeja

Zavesky

et al. 2013

SPIE Proceedings

View full text Add to dashboard Cite

Section: Related Workmentioning

confidence: 99%

VidCat: an image and video analysis service for personal media management

Begeja

Zavesky

et al. 2013

SPIE Proceedings

View full text Add to dashboard Cite

“…For example, Fitzgibbon et al [3], [5] proposed an affine invariant distance metric which is robust to a desired group of transformations for video face clustering. Huang et al [14] proposed to cluster faces with multi-views in a video sequence. They clustered video faces based on pose grouping results.…”

Section: Related Workmentioning

confidence: 99%

“…We further improve the approach into the multi-view framework, named Constrained Multi-view Video Face Clustering (CMVFC). To distinguish our method from the method [14] which defines view in a geometry point of view, we define the multi-view face clustering of our interest as follow:…”

Section: E Constrained Multi-view Spectral Clusteringmentioning

confidence: 99%

Constrained Multi-View Video Face Clustering

Cao

Zhang

Zhou

et al. 2015

IEEE Trans. on Image Process.

105

View full text Add to dashboard Cite

In this paper, we focus on face clustering in videos. To promote the performance of video clustering by multiple intrinsic cues, i.e., pairwise constraints and multiple views, we propose a constrained multi-view video face clustering method under a unified graph-based model. First, unlike most existing video face clustering methods which only employ these constraints in the clustering step, we strengthen the pairwise constraints through the whole video face clustering framework, both in sparse subspace representation and spectral clustering. In the constrained sparse subspace representation, the sparse representation is forced to explore unknown relationships. In the constrained spectral clustering, the constraints are used to guide for learning more reasonable new representations. Second, our method considers both the video face pairwise constraints as well as the multi-view consistence simultaneously. In particular, the graph regularization enforces the pairwise constraints to be respected and the co-regularization penalizes the disagreement among different graphs of multiple views. Experiments on three real-world video benchmark data sets demonstrate the significant improvements of our method over the state-of-the-art methods.

show abstract

“…[2] propose using a dissimilarity matrix from pre-existing face clusters; Fitzsgibbon and Zisserman [3] use joint manifold distance; Tao and Tan propose dividing face sequences by pose and then applying additional constraints obtained from domain knowledge [4]. There are other methods that have been proposed in [5], [6], [7].…”

Section: Introductionmentioning

confidence: 99%