Image set-based collaborative representation for face recognition

$59

Description

We proposed a novel image set based collaborative representation and classification (ISCRC) scheme for image set based face recognition (ISFR).

The query set was modeled as a convex or regularized hull, and a collaborative representation based set to sets distance (CRSSD) was defined by representing the hull of query set over all the gallery sets.

The CRSSD considers the correlation and distinction of sample images within the query set and the relationship between the gallery sets. With CRSSD, the representation residual of the hull of query set by each gallery set can be computed and used for classification.

Experiments on the three benchmark ISFR databases showed that the proposed ISCRC is superior to state-of-the-art ISFR methods in terms of both recognition rates and efficiency.

 

Image set based collaborative representation and classification

 

 

blank

Recognition performance of RH-ISCRC-l2 on CMU MoBo with different 1 and 2. Different colors represent different accuracy. Top: 50 frames per set; middle: 100 frames per set; bottom: 200 frames per set.

 

 

 

ref :

Zhu, Pengfei, Wangmeng Zuo, Lei Zhang, Simon Chi-Keung Shiu, and David Zhang. “Image set-based collaborative representation for face recognition.” IEEE transactions on information forensics and security 9, no. 7 (2014): 1120-1132.

Relaxed Collaborative Representation for Patter Classification

Sparse or Collaborative Representation for Face Recognition

Reviews

There are no reviews yet.

Be the first to review “Image set-based collaborative representation for face recognition”