簡易檢索 / 詳目顯示

回結果列表
研究生： 陳品諼
Chen, Pin-Hsuan
論文名稱： 多變形殘差神經網絡與特徵鑑別器之濕指紋去噪
Deformed Residual Neural Network with Featured Discriminator for Wet Fingerprint denoising
指導教授： 邱瀞德
Chiu, Ching-Te
口試委員: 林輝堂
Lin, Hui-Tang
徐茂修
Hsu, Mao-Hsiu
學位類別： 碩士
Master
系所名稱： 電機資訊學院 - 通訊工程研究所
Communications Engineering
論文出版年： 2023
畢業學年度： 112
語文別： 英文
論文頁數： 60
中文關鍵詞： 指紋身份驗證 、指紋識別 、圖像增強 、濕指紋 、圖像生成器 、卷積神經網絡 、鑑別器 、殘差神經網路
外文關鍵詞： Fingerprint authentication, Fingerprint recognition, Image enhancement, Wet fingerprints, Image generator, Convolution Neural Net Work, Discriminator, Residual Neural Net Work
相關次數： 點閱：56  下載：0 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報

    • 指紋辨識是生物辨識身份驗證系統中的關鍵組件，對於保障設備和個人數據的安全起著至關重要的作用。儘管電容式感測器因其緊湊的設計和能夠整合到手機中而變得越來越受歡迎，但它們容易受到噪聲和環境因素（如汗水）的影響。當處理從電容式感測器中獲得的濕指紋去噪時，常常會面臨一些挑戰，例如指紋面積較小、特徵數量不足，以及由水造成的廣泛黑暗區域。
    本文提出了一種用於從電容式感測器獲得的濕指紋去噪的方法。我們提出的方法稱為“DRB-FD＂，它結合了 Featured Discriminator（FD）和Deformed Residual Block（DRB）的功能，並整合了關注機制（CBAM）、Drop-out 層和預激活。FD 從指紋中提取重要的特徵並分配特徵權重，以優先處理圖像區域，增強模型對關鍵區域的關注。DRB 優化了殘差塊的架構，以捕捉細節並減少過擬合，確保了穩健的去噪性能。
    實驗結果顯示，DRB-FD 模型在去噪電容式感測器指紋方面取得了顯著的改善。在 Nasic9395_0606_aug 電容式指紋數據集中，與原本標準PGT-Net 相比，它實現了 73.1％的顯著進步（FRR 減少了 25％），在其中FD 提升 47.4％（FRR 減少了16.2％），DRB 則是提升了 8.9％（FRR 減少了8.8％）。
    本研究提出了 DRB-FD，用於從電容式感測器獲得的指紋進行去噪。通過利用 Feature Discriminator(FD) 和 Deformed Residual Block(DRB)，我們的方法有顯著的 FRR 改善，證明了它針對濕指紋除噪的準確性和可靠性。

    Fingerprint recognition is a pivotal component of biometric authentication systems, playing a crucial role in securing devices and personal data. While capacitive sensors have gained popularity for their compact design and integration into mobile devices, they are susceptible to noise and environmental factors(sweat). When dealing with wet fingerprint denoising from capacitive sensors, challenges often arise due to factors such as small fingerprint areas, insufficient feature quantity, and extensive dark regions caused by water. This paper presents an advanced approach to denoising wet fingerprint data acquired through capacitive sensors.
    Our proposed method called ”DRB-FD” which combines the power of a Featured Discriminator (FD) and a Deformed Residual Block (DRB), incorporating
    attention mechanisms(CBAM), drop-out layers, and pre-activation. The FD extracts essential fingerprint features and assigns feature weights to prioritize image regions, enhancing the model’s focus on critical areas. The DRB optimizes the residual block architecture to capture fine details and mitigate overfitting, ensuring robust denoising performance.
    Experimental results that FDB-FD model demonstrate substantial improvements in denoising capacitive sensor fingerprints. In the Nasic9395_0606_aug dataset, it
    achieved a remarkable 73.1% improvement(FRR decreased by 25%) compared to the baseline PGT-Net. The FD and DRB contributed significantly, with improvements of 47.4%(FRR decreased by 16.2%) and 48.9%(FRR decreased by 8.8%), respectively.
    In conclusion, this study presents an innovative approach to denoising finger print data acquired through capacitive sensors. By leveraging FD, DRB, and feature analysis, our method achieves substantial FRR improvements, demonstrating its effectiveness in enhancing fingerprint recognition systems’ accuracy and reliability.

    Acknowledgements 摘要 i Abstract ii 1 Introduction 1 1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.3 Problem Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.4 Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2 Related Work 9 2.1 Residual Neural Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.2 PGT−Net . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.2.1 Data Flow of PGT-Net . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.2.2 PGT-Net-block-84 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.3 Discriminator−VGG19 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 2.4 Attention Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2.4.1 Squeeze-and-Excitation Block (SE-Block) . . . . . . . . . . . . . . . . 19 2.4.2 Convolutional Block Attention Module(CBAM) . . . . . . . . . . . . 21 2.5 WRNs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 2.6 Full Pre-Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3 DRB-FD 25 3.1 Featured Discriminator(FD) . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.1.1 Discriminator−VGG19 . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.1.2 Feature Weight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 3.1.3 Identity Preserving Loss . . . . . . . . . . . . . . . . . . . . . . . . . 29 3.2 Deformed Residual Block(DRB) . . . . . . . . . . . . . . . . . . . . . . . . . 30 3.2.1 Attention Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 3.2.2 WRNs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 3.2.3 Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 3.3 Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 4 Experiment 39 4.1 Dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 4.1.1 Training Dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 4.1.2 Nasic9395_testset . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 v 4.2 Featured Discriminator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 4.2.1 Discriminator-VGG19 . . . . . . . . . . . . . . . . . . . . . . . . . . 41 4.2.2 Feature weight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 4.2.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 4.3 Deformed Residual Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 4.3.1 Attention Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 4.3.2 Drop Out . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 4.3.3 Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 5 Conclusion 55 References 57

    [1] S. R. J. S. Kaiming He, Xiangyu Zhang, “Deep residual learning for image recognition,”Computer Vision and Pattern Recognition(CVPR), 2015.
    [2] A.-T. H. M.-H. H. L. W. J.-M. H. Yu-Ting Li, Ching-Te Chiu, “Pgt-net: Progressive guided multi-task neural network for small-area wet fingerprint denoising and recognition,” 2023.
    [3] W. X. B. L. X. L. Jie Song, Huawei Yi, “Gram-gan: Image super-resolution based on gram matrix and discriminator perceptual loss,” Multidisciplinary Digital Publishing Institute(MDPI), 2023.
    [4] S. R. J. S. Kaiming He, Xiangyu Zhang, “Identity mappings in deep residual networks,” European Conference on Computer Vision (ECCV), 2016.
    [5] A. Z. Karen Simonyan, “Very deep convolutional networks for large-scale image recognition,” 2014 ImageNet Large Scale Visual Recognition Challenge (ILSVRC), 2014.
    [6] M. Kücken and A. Newell, “Fingerprint formation,” Journal of theoretical biology, vol.235, pp. 71–83, 08 2005.
    [7] A. Jain, S. Prabhakar, and S. Pankanti, “On the similarity of identical twin fingerprints,” Pattern Recognition, vol. 35, pp. 2653–2663, 11 2002.
    [8] J. Li, J. Feng, and C.-C. J. Kuo, “Deep convolutional neural network for latent fingerprint enhancement,” Signal Processing: Image Communication, vol. 60, pp.52–63, 2018. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S0923596517301492
    [9] J. Yang, N. Xiong, and A. V. Vasilakos, “Two-stage enhancement scheme for low-quality fingerprint images by learning from the images,” IEEE Transactions on Human-Machine Systems, vol. 43, no. 2, pp. 235–248, 2013.
    [10] S. Ali, M. Khan, and A. Aslam, “Fingerprint matching, spoof and liveness detection: classification and literature review,” Frontiers of Computer Science, vol. 15, p. 151310, 092020.
    [11] Y. Li, Q. Xia, C. Lee, S. Kim, and J. Kim, “A robust and efficient fingerprint image restoration method based on a phase-field model,” Pattern Recognition, vol. 123, p.108405, 2022. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S0031320321005811
    [12] F. u. A. A. Minhas, M. Arif, and M. Hussain, “Fingerprint identification and verification system using minutiae matching,” 01 2004.
    [13] S. Sudiro and R. Trisno, “Adaptable fingerprint minutiae extraction algorithm based-on crossing number method for hardware implementation using fpga device,” International Journal of Computer Science, Engineering and Information Technology, vol. 2, 06 2012.
    [14] B. Bhanu and X. Tan, “Fingerprint indexing based on novel features of minutiae triplets,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, no. 5, pp. 616–622, 2003.
    [15] Q. Zhao, D. Zhang, L. Zhang, and N. Luo, “High resolution partial fingerprint alignment using pore–valley descriptors,” Pattern Recognition, vol. 43, no. 3, pp.1050–1061, 2010. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S0031320309003045
    [16] J. Jaam, M.-L. Rebaiaia, and A. Hasnah, “A fingerprint minutiae recognition system based on genetic algorithms,” Int. Arab J. Inf. Technol., vol. 3, pp. 242–248, 01 2006.
    [17] Q. Zhao, L. Zhang, D. Zhang, and N. Luo, “Direct pore matching for fingerprint recognition,” in Advances in Biometrics, M. Tistarelli and M. S. Nixon, Eds. Berlin, Heidelberg:Springer Berlin Heidelberg, 2009, pp. 597–606.
    [18] M. A. Medina-Pérez, M. García-Borroto, A. E. Gutierrez-Rodríguez, and L. AltamiranoRobles, “Improving fingerprint verification using minutiae triplets,” Sensors, vol. 12,no. 3, pp. 3418–3437, 2012. [Online]. Available: https://www.mdpi.com/1424-8220/12/3/3418
    [19] Z. Chen and C. Kuo, “A topology-based matching algorithm for fingerprint authentication,” in Proceedings. 25th Annual 1991 IEEE International Carnahan Conference on Security Technology, 1991, pp. 84–87.
    [20] T. Kristensen, “Two different regimes of fingerprint identification–a comparison,” American Journal of Computational and Applied Mathematics, vol. 2, pp. 1–9, 05 2012.
    [21] B. Tan, A. Lewicke, D. Yambay, and S. Schuckers, “The effect of environmental conditions and novel spoofing methods on fingerprint anti-spoofing algorithms,” in 2010 IEEE International Workshop on Information Forensics and Security, 2010, pp. 1–6.
    [22] A. Jain, S. Prabhakar, and L. Hong, “A multichannel approach to fingerprint classification,” Pattern Analysis and Machine Intelligence, IEEE Transactions on, vol. 21, pp. 348 – 359, 05 1999.
    [23] M.-T. Leung, W. Engeler, and P. Frank, “Fingerprint image processing using neural networks,” in IEEE TENCON’90: 1990 IEEE Region 10 Conference on Computer and Communication Systems. Conference Proceedings, 1990, pp. 582–586 vol.2.
    [24] E. Liu, H. Zhao, F. Guo, J. Liang, and J. Tian, “Fingerprint segmentation based on an adaboost classifier,” Frontiers of Computer Science in China, vol. 5, pp. 148–157, 062011.
    [25] T. Kristensen, J. Borthen, and K. Fyllingsnes, “Comparison of neural network based fingerprint classification techniques,” 08 2007, pp. 1043–1048.
    [26] A. Rattani and A. Ross, “Automatic adaptation of fingerprint liveness detector to new spoof materials,” in IEEE International Joint Conference on Biometrics, 2014, pp. 1–8.
    [27] Z. Jin, A. Teoh, T. S. Ong, and T. Connie, “Secure minutiae-based fingerprint templates using random triangle hashing,” vol. 5857, 11 2009, pp. 521–531.
    [28] S. Minaee and Y. Wang, “Fingerprint recognition using translation invariant scattering network,” CoRR, vol. abs/1509.03542, 2015. [Online]. Available: http://arxiv.org/abs/1509.03542
    [29] Y. Ding, A. Rattani, and A. Ross, “Bayesian belief models for integrating match scores with liveness and quality measures in a fingerprint verification system,” in 2016 International Conference on Biometrics (ICB), 2016, pp. 1–8.
    [30] X. Z. J. S. Liangyu Chen, Xiaojie Chu, “Simple baselines for image restoration,” European Conference on Computer Vision(ECCV), 2022.
    [31] S.-W. K. J.-P. H. S.-J. B. S.-W. J. S.-J. K. Seo-Won Ji, Jeongmin Lee, “Xydeblur: Divide and conquer for single image deblurring,” IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022.
    [32] S. S.-K. M. L. Reyhaneh Neshatavar, Mohsen Yavartanoo, “Cvf-sid: Cyclic multi-variate function for self-supervised image denoising by disentangling noise from image,” IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022.
    [33] T. K. W. Sirojbek Safarov, “A-denseunet: Adaptive densely connected unet for polypsegmentation in colonoscopy images with atrous convolution,” Multidisciplinary Digital Publishing Institute(MDPI), 2021.
    [34] K.-H. L. Chi-Mao Fan, Tsung-Jung Liu, “Selective residual m-net for real image denoising,” European Signal Processing Conference (EUSIPCO), 2022.
    [35] J. F. Y. L. Yao Tang, Fei Gao, “Fingernet: An unified deep network for fingerprint minutiae extraction,” IEEE International Joint Conference on Biometrics (IJCB), 2017.
    [36] J. S. Sukesh Adiga V, “Fpd-m-net: Fingerprint image denoising and inpainting using m-net based convolutional neural networks,” European Conference on Computer Vision(ECCV),2018.
    [37] T. B. Olaf Ronneberger, Philipp Fischer, “U-net: Convolutional networks for biomedical image segmentation,” Medical Image Computing and Computer Assisted Intervention Society(MICCAI), 2015.
    [38] D. L. W. Kalaivani Sundararajan, “Deep learning for biometrics: A survey,” in ACM Computing Surveys, vol. 51, no. 65, 2018, pp. 1–34. [Online]. Available:https://doi.org/10.1145/3190618
    [39] A. A. A. A. T.-V. B. T. M. Bhavesh Pandya, Georgina Cosma, “Fingerprint classification using a deep convolutional neural network,” 4th International Conference on Information Management (ICIM), 2018.
    [40] P. S. T. J. K. C. Y.-H. W. Xin Mao, Zhaoyu Su, “Is discriminator a good feature extractor?”2019.
    [41] Y. Tao, “Image style transfer based on vgg neural network model,” IEEE International Conference on Advances in Electrical Engineering and Computer Applications (AEECA),2022.
    [42] M. B. Leon A. Gatys, Alexander S. Ecker, “Image style transfer using convolutional neural networks,” IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016.
    [43] N. V. Pei Wang, Yijun Li, “Rethinking and improving the robustness of image style transfer,” IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2021.
    [44] G. S. Jie Hu, Li Shen, “Squeeze-and-excitation networks,” 2018 Conference on Computer Vision and Pattern Recognition (CVPR), 2018.
    [45] J.-Y. L. I. S. K. Sanghyun Woo, Jongchan Park, “Cbam: Convolutional block attention module,” European Conference on Computer Vision (ECCV), 2018.
    [46] N. K. Sergey Zagoruyko, “Wide residual networks,” 2016 The British Machine Vision Conference (BMVC), 2016.

    QR CODE