Deep Learning for Face Anti-Spoofing: An End-to-End Approach
標籤： anti-spoofing
論文出處： IEEE 2017
本文提出的方法
本文提出的是一種端到端的CNN架構，因爲以前提出的方法，即使是基於CNN的，但是隻是把CNN作爲了一種特徵提取器，而之後再把特徵使用SVM進行分類，而本文提出的方案，是一種端到端的方案，末端接了全連接並且使用softmax進行分類。並且給出了一些網絡調節的小技巧比如50RS-30SeC-1E等，這些方法如果將來我要是做基於CNN的則可以參考。
網絡是VGG以及兩個基於VGG的延伸網絡，而對於數據的預處理，只是先把人臉給提取出來，並沒有做其他處理。
除此之外，本文給出了很多種結果比較方法，比如top-1 percent accuracy或者accuracy using threshold-operation等，並且與其他方案做了比較。發現還是有所提升。
收穫
1、一種端到端的結構進行欺詐檢測
金句： The remarkable success of Convolutional Neural Networks
(CNN) [8] in ImageNet [6] competition has attracted a multitude of researchers in the computer vision community to
investigate its potential latent capabilities in attaining such
a high performance.
參考文獻重點摘錄可作爲以後讀
其他CNN方法
[20] D. Gragnaniello, C. Sansone, G. Poggi, and L. Verdoliva, “Biometric
spoofing detection by a domain-aware convolutional neural network,” in
Signal-Image Technology & Internet-Based Systems (SITIS), 2016 12th
International Conference on. IEEE, 2016, pp. 193–198.
[21] A. Alotaibi and A. Mahmood, “Deep face liveness detection based on
nonlinear diffusion using convolution neural network,” Signal, Image
and Video Processing, pp. 1–8, 2016.
[22] ——, “Enhancing computer vision to detect face spoofing attack utilizing a single frame from a replay video attack using deep learning,”
in Optoelectronics and Image Processing (ICOIP), 2016 International
Conference on. IEEE, 2016, pp. 1–5.
[23] J. Yang, Z. Lei, and S. Z. Li, “Learn convolutional neural network for
face anti-spoofing,” arXiv preprint arXiv:1408.5601, 2014.
[24] L. Li, X. Feng, Z. Boulkenafet, Z. Xia, M. Li, and A. Hadid, “An original
face anti-spoofing approach using partial convolutional neural network,”
in Image Processing Theory Tools and Applications (IPTA), 2016 6th
International Conference on. IEEE, 2016, pp. 1–6