TY - GEN
T1 - High-Resolution Face Recognition Via Deep Pore-Feature Matching
AU - Lai, Shun Cheung
AU - Kong, Minna
AU - Lam, Kin Man
AU - Li, Dong
PY - 2019/9/22
Y1 - 2019/9/22
N2 - Because of the advancement of capturing devices, both image resolution and image quality have been significantly improved. Efficiently utilizing facial information is beneficial in enhancing the performance of face recognition methods. For high-resolution face images, pore-scale facial features can be observed. The positions and local patterns of pore features are biologically discriminative, so they can be explored for face identification. In this paper, we extend the previous work on pore-scale features, by proposing a new learning-based descriptor, namely PoreNet. Experiment results show that our proposed descriptor achieves an excellent performance on two high-resolution face datasets, namely Bosphorus and MultiPIE. More importantly, our proposed method significantly outperforms the state-of-the-art Convolutional Neural Network (CNN)-based face recognition method, when query faces are highly occluded. The code of our proposed method is available at: https://github.com/johnnysclai/PoreNet.
AB - Because of the advancement of capturing devices, both image resolution and image quality have been significantly improved. Efficiently utilizing facial information is beneficial in enhancing the performance of face recognition methods. For high-resolution face images, pore-scale facial features can be observed. The positions and local patterns of pore features are biologically discriminative, so they can be explored for face identification. In this paper, we extend the previous work on pore-scale features, by proposing a new learning-based descriptor, namely PoreNet. Experiment results show that our proposed descriptor achieves an excellent performance on two high-resolution face datasets, namely Bosphorus and MultiPIE. More importantly, our proposed method significantly outperforms the state-of-the-art Convolutional Neural Network (CNN)-based face recognition method, when query faces are highly occluded. The code of our proposed method is available at: https://github.com/johnnysclai/PoreNet.
KW - Face recognition
KW - feature extraction
KW - high-resolution face recognition
KW - pore-scale facial feature
UR - http://www.scopus.com/inward/record.url?scp=85076808677&partnerID=8YFLogxK
U2 - 10.1109/ICIP.2019.8803686
DO - 10.1109/ICIP.2019.8803686
M3 - Conference article published in proceeding or book
AN - SCOPUS:85076808677
T3 - Proceedings - International Conference on Image Processing, ICIP
SP - 3477
EP - 3481
BT - 2019 IEEE International Conference on Image Processing, ICIP 2019 - Proceedings
PB - IEEE Computer Society
T2 - 26th IEEE International Conference on Image Processing, ICIP 2019
Y2 - 22 September 2019 through 25 September 2019
ER -