Face Clustering Utilizing Scalable Sparse Subspace Clustering And The Image Gradient Feature Descriptor

EasyChair Preprint no. 118

Abstract

Face clustering is an important topic in  computer vision. It aims to put together facial images that belong to the same person. Spectral clustering-based algorithms are often used for accurate face clustering. However, a big occlusion matrix is usually needed to deal with the noise and sparse outlier terms, which makes the sparse coding process computationally expensive. Thus spectral clustering-based algorithms are difficult to extend to large scale datasets. In this paper, we use the image gradient feature descriptor and scalable Sparse Subspace Clustering algorithm for large scale and high accuracy face clustering. Within the image gradient feature descriptor, the scalable Sparse Subspace Clustering algorithm can be used in large scale face datasets without sacrificing clustering performance. Experimental results show that our algorithm is robust to illumination, occlusion, and achieves a relatively high clustering accuracy on the Extended Yale B and AR datasets.

Keyphrases: face clustering, image gradient feature descriptor, scalable Sparse Subspace Clustering

@Booklet{EasyChair:118,
  author = {Mingkang Liu and Qi Li and Zhenan Sun and Hongwen Zhang and Qiyao Deng},
  title = {Face Clustering Utilizing Scalable Sparse Subspace Clustering And The Image Gradient Feature Descriptor},
  howpublished = {EasyChair Preprint no. 118},
  doi = {10.29007/hgz2},
  year = {EasyChair, 2018}}