臺灣博碩士論文加值系統

傳統的人臉偵測及辨識技術容易受到光線、遮蔽、背景等因素影響，導致辨識率
無法提升。近年來，諸多文獻提出了許多提高人臉辨識精準度的深度學習技術，隨者
圖形處理器的發展，這些深度學習技術得以實現即時應用。然而這些深度學習技術都
是獨立運行，一個完整的系統需要將這些深度學習技術整合，因此本論文設計了一個
系統整合GUI介面，整合了多執行緒即時影像擷取及顯示模組(IP cam/Web cam)、參數
修訂、人員管理、深度學習技術人臉偵測技術MTCNN與深度學習技術人臉辨識技術
FaceNet。為了驗證本論文整合系統的可行性，我們收集了70位人員的人臉影像以及30
位人員的人臉影片，作為實際使用的測試。所提出之整合系統使用中低階圖形處理器
(GTX1050/GTX1660)即可達到99.3%的準確率以及20-25FPS的即時運算。

The traditional face detection and recognition technology is easily affected by factors such as light, shadow, and background, resulting in an inability to improve the recognition rate. In recent years, many literatures have proposed a number of deep learning techniques to improve the accuracy of face recognition. With the development of graphics processors, these deep learning techniques can be applied in real time. Based on the system development perspective, these deep learning technologies are all independent. A complete system needs to integrate these deep learning technologies. This paper designs a system integration GUI interface and integrates multi-threaded instant image capture and display modules, parameter revision, personnel management, MTCNN face detection technology and FaceNet face recognition technology. In order to verify the feasibility of the integrated system, we collected the face images of 70 people and the face videos of 30 people as the actual test. The use of this integrated system uses a low-end graphics processor (GTX1050/GTX1660) to achieve 99.3% accuracy and 20-25FPS real-time operation.

摘要 i
Abstract ii
致謝 iii
目錄 iv
圖目錄 v
第一章 簡介 1
1.1 研究背景與動機 1
1.2 相關文獻探討 1
1.3 研究貢獻 3
1.4 系統架構 3
第二章 人臉偵測 5
2.1 MTCNN簡介 5
2.2 MTCNN Loss function 7
2.3 多角度人臉偵測實驗結果 8
第三章 人臉辨識 9
3.1 FaceNet深度學習網路架構 9
3.2 FaceNet Loss function 12
3.3 人臉辨識實驗結果 13
第四章 智慧人臉辨識系統整合 15
第五章 實驗及實驗結果 21
第六章 結論與未來展望 25
參考文獻 26

[1]K. Zhang, Z. Zhang, Z. Li, Y. Qiao, “Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks,” IEEE Signal Processing Letters (SPL), vol. 23, no. 10, pp. 1499-1503, 2016.
[2]F. Schroff, D. Kalenichenko and J. Philbin, “FaceNet: A Unified Embedding for Face Recognition and Clustering,” IEEE Conference on Computer Vision and Pattern Recognition, pp. 815-523, 2015.
[3]P. Viola, M. Jones, and D. Snow, “Detecting pedestrians using patterns of motion and appearance,” IEEE International Conference on Computer Vision, vol. 2, pp. 734-741, 2003.
[4]D. Lowe, “Distinctive image features from scale-invariant keypoints,” International Journal of Computer Visio, pp. 91-110, 2004.
[5]N. Dalal and B. Triggs, “Histograms of oriented gradients for human detection,” IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 1, pp. 886-893, 2005.
[6]P. Dollar, Z. Tu, P. Perona, and S. Belongie, “Integral channel features,” British Machine Vision Conference, pp.91.1-91.11, 2009.
[7]P. Dollar, R. Appel, and W. Kienzle, “Crosstalk cascades for frame-rate pedestrian detection,” European Conference on Computer Vision, pp. 645-659, 2012.
[8]Z. Cao, Q. Yin, X. Tang, and J. Sun. “Face recognition with learning based descriptor,” IEEE Conference on Computer Vision and Pattern Recognition, pp. 2707–2714, 2010.
[9]T.-H. Chan, K. Jia, S. Gao, J. Lu, Z. Zeng, and Y. Ma. “Pcanet: A simple deep learning baseline for image classification?,” IEEE Transactions on Image Processing, vol. 24, no. 12, pp. 5017–5032, 2015.
[10]S. Yang, P. Luo, C. C. Loy, and X. Tang. “WIDER FACE: A face detection benchmark,” IEEE Conference on Computer Vision and Pattern Recognition, pp. 5525-5533, 2016.
[11]Jain and E. Learned-Miller. “FDDB: A benchmark for face detection in unconstrained settings,” Technical Report UM-CS-2010-009, University of Massachusetts, Amherst, 2010.
[12]H. Jiang and E. Learned-Miller, “Face detection with the faster R-CNN,” IEEE International Conference on Automatic Face & Gesture Recognition, pp. 650-657, 2017.
[13]S. Yang, P. Luo, C. C. Loy, and X. Tang, “Faceness-Net: Face Detection through Deep Facial Part Responses,“ IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 40, no. 8, pp. 1845-1859, 2017.
[14]K. Simonyan and A. Zisserman. “Very deep convolutional networks for large-scale image recognition,” International Conference on Learning Representations, 2015.
[15]Y. Taigman, M. Yang, M. Ranzato, and L. Wolf. “Deepface: Closing the gap to human-level performance in face verification,” IEEE Conference on Computer Vision and Pattern Recognition, pp. 1701–1708, 2014.
[16]Y. Sun, L. Ding, X. Wang, and X. Tang. “Deepid3: Face recognition with very deep neural networks,” CoRR, abs/1502.00873, 2015
[17]G. B. Huang, M. Ramesh, T. Berg, and E. Learned-Miller. “Labeled Faces in the Wild: A Database for Studying Face Recognition in Unconstrained Environments,” University of Massachusetts, Amherst, Technical Report, pp.07-49, 2007.