TY - GEN
T1 - The do's and don'ts for CNN-Based face verification
AU - Bansal, Ankan
AU - Castillo, Carlos
AU - Ranjan, Rajeev
AU - Chellappa, Rama
N1 - Funding Information:
This research is based upon work supported by the Office of the Director of National Intelligence (ODNI), Intelligence Advanced Research Projects Activity (IARPA), via IARPA R&D Contract No. 2014-14071600012. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of the ODNI, IARPA, or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright annotation thereon.
Publisher Copyright:
© 2017 IEEE.
PY - 2017/7/1
Y1 - 2017/7/1
N2 - While the research community appears to have developed a consensus on the methods of acquiring annotated data, design and training of CNNs, many questions still remain to be answered. In this paper, we explore the following questions that are critical to face recognition research: (i) Can we train on still images and expect the systems to work on videos? (ii) Are deeper datasets better than wider datasets? (iii) Does adding label noise lead to improvement in performance of deep networks? (iv) Is alignment needed for face recognition? We address these questions by training CNNs using CASIA-WebFace, UMD-Faces, and a new video dataset and testing on YouTube-Faces, IJB-A and a disjoint portion of UMDFaces datasets. Our new data set, which will be made publicly available, has 22,075 videos and 3,735,476 human annotated frames extracted from them.
AB - While the research community appears to have developed a consensus on the methods of acquiring annotated data, design and training of CNNs, many questions still remain to be answered. In this paper, we explore the following questions that are critical to face recognition research: (i) Can we train on still images and expect the systems to work on videos? (ii) Are deeper datasets better than wider datasets? (iii) Does adding label noise lead to improvement in performance of deep networks? (iv) Is alignment needed for face recognition? We address these questions by training CNNs using CASIA-WebFace, UMD-Faces, and a new video dataset and testing on YouTube-Faces, IJB-A and a disjoint portion of UMDFaces datasets. Our new data set, which will be made publicly available, has 22,075 videos and 3,735,476 human annotated frames extracted from them.
UR - https://www.scopus.com/pages/publications/85046264707
UR - https://www.scopus.com/pages/publications/85046264707#tab=citedBy
U2 - 10.1109/ICCVW.2017.299
DO - 10.1109/ICCVW.2017.299
M3 - Conference contribution
AN - SCOPUS:85046264707
T3 - Proceedings - 2017 IEEE International Conference on Computer Vision Workshops, ICCVW 2017
SP - 2545
EP - 2554
BT - Proceedings - 2017 IEEE International Conference on Computer Vision Workshops, ICCVW 2017
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 16th IEEE International Conference on Computer Vision Workshops, ICCVW 2017
Y2 - 22 October 2017 through 29 October 2017
ER -