Self-Supervised Learning Framework toward State-of-the-Art Iris Image Segmentation

被引:4
作者
Putri, Wenny Ramadha [1 ]
Liu, Shen-Hsuan [1 ]
Aslam, Muhammad Saqlain [1 ]
Li, Yung-Hui [2 ]
Chang, Chin-Chen [3 ]
Wang, Jia-Ching [1 ]
机构
[1] Natl Ctr Univ, Dept Comp Sci & Informat Engn, Taoyuan 32001, Taiwan
[2] Hon Hai Res Inst, AI Res Ctr, Taipei 114699, Taiwan
[3] Feng Chia Univ, Dept Informat Engn & Comp Sci, Taichung 40724, Taiwan
关键词
data augmentation; iris segmentation; generative adversarial network; image semantic segmentation; biometrics; RECOGNITION; NETWORK; NET;
D O I
10.3390/s22062133
中图分类号
O65 [分析化学];
学科分类号
070302 ; 081704 ;
摘要
Iris segmentation plays a pivotal role in the iris recognition system. The deep learning technique developed in recent years has gradually been applied to iris recognition techniques. As we all know, applying deep learning techniques requires a large number of data sets with high-quality manual labels. The larger the amount of data, the better the algorithm performs. In this paper, we propose a self-supervised framework utilizing the pix2pix conditional adversarial network for generating unlimited diversified iris images. Then, the generated iris images are used to train the iris segmentation network to achieve state-of-the-art performance. We also propose an algorithm to generate iris masks based on 11 tunable parameters, which can be generated randomly. Such a framework can generate an unlimited amount of photo-realistic training data for down-stream tasks. Experimental results demonstrate that the proposed framework achieved promising results in all commonly used metrics. The proposed framework can be easily generalized to any object segmentation task with a simple fine-tuning of the mask generation algorithm.
引用
收藏
页数:25
相关论文
共 76 条
[1]  
Al-Raisi Ahmad N., 2008, Telematics and Informatics, V25, P117, DOI 10.1016/j.tele.2006.06.005
[2]  
[Anonymous], 2015, 151105644 ARXIV
[3]  
[Anonymous], 2015, P IEEE C COMP VIS PA
[4]   FRED-Net: Fully residual encoder-decoder network for accurate iris segmentation [J].
Arsalan, Muhammad ;
Kim, Dong Seop ;
Lee, Min Beom ;
Owais, Muhammad ;
Park, Kang Ryoung .
EXPERT SYSTEMS WITH APPLICATIONS, 2019, 122 :217-241
[5]   IrisDenseNet: Robust Iris Segmentation Using Densely Connected Fully Convolutional Networks in the Images by Visible Light and Near-Infrared Light Camera Sensors [J].
Arsalan, Muhammad ;
Naqvi, Rizwan Ali ;
Kim, Dong Seop ;
Nguyen, Phong Ha ;
Owais, Muhammad ;
Park, Kang Ryoung .
SENSORS, 2018, 18 (05)
[6]   Iris Segmentation Using Geodesic Active Contours and GrabCut [J].
Banerjee, Sandipan ;
Mery, Domingo .
IMAGE AND VIDEO TECHNOLOGY - PSIVT 2015 WORKSHOPS, 2016, 9555 :48-60
[7]   An end to end Deep Neural Network for iris segmentation in unconstrained scenarios [J].
Bazrafkan, Shabab ;
Thavalengal, Shejin ;
Corcoran, Peter .
NEURAL NETWORKS, 2018, 106 :79-95
[8]   Inside-Outside Net: Detecting Objects in Context with Skip Pooling and Recurrent Neural Networks [J].
Bell, Sean ;
Zitnick, C. Lawrence ;
Bala, Kavita ;
Girshick, Ross .
2016 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2016, :2874-2883
[9]  
Chen L.C., 2015, ICLR
[10]   DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs [J].
Chen, Liang-Chieh ;
Papandreou, George ;
Kokkinos, Iasonas ;
Murphy, Kevin ;
Yuille, Alan L. .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2018, 40 (04) :834-848