Face Verification results on VADANA dataset
People: Gowri Somanath, Rohith MV, Chandra Kambhamettu
Experiments for age gap [0,2]
The following table reports the accuracy at Equal Error Rate (EER), where EER is defined as the error when Correct Acceptance Rate (CAR) = Correct Rejection Rate (CRR).
CAR= (# of correctly classified intra-pairs) / (Total # of intra-pairs)
CRR= (# of correctly classified extra-pairs) / (Total # of extra-pairs)
See paper for details on experiments.
Algorithm,Dataset | VADANA | FGNET |
Baseline: Eigenfaces (PCA+RF) | 52.33 | 99.33 |
Nowak pair matching (SIFT+ERFC) Nowak | 61.52 | 67.0 +- 2.2 |
Ling et. al (GOP+SVM) Ling | 57.43 | 73.0 |
Code
The matlab source code used for the baseline method and our implementation of the GOP+SVM method can be found here.
References
Nowak E. Nowak and F. Jurie. Learning visual similarity measures for comparing never seen objects. In Conference on Computer Vision & Pattern Recognition, 2007.
Ling H. Ling, S. Soatto, N. Ramanathan, and D. Jacobs. Face verification across age progression using discriminative methods. Information Forensics and Security, IEEE Transactions on, 5(1):82 –91, March 2010.