Biometric spoofing detection by a Domain-aware Convolutional Neural Network

Biometric spoofing detection by a Domain-aware Convolutional Neural Network

标签: antispoofing 论文


论文出处:2016 12th International Conference on Signal-Image Technology & Internet-Based Systems

基于经验信息

现在很多地方都用CNN网络进行预测,但是有些地方效果很好,但是有些效果不好,因为在CNN上面学习到的东西不可见,并且可能有些东西并不能自助学习到,比如在[13]的一个例子,CNN就学习不到中值滤波这样简单的滤波器,所以CNN网络在有些时候需要加一些限制或者约束,我们根据我们现有的知识,让我们的CNN网络向着我们觉得正确的方向去学习,这也就是基于经验的CNN网络,加入了人为的因素。而本文的思想正是基于此。提出了从图片中学习到residual image,而为了学习到这种信息,所以在损失函数中加了正则项,如下:

image

这样的话就可以引导着CNN进行学习了。
本文的网络结构就是普通CNN结构。在数据处理上面进行了图片旋转90°,所以图片多了4倍,并且还着重提出,不要进行resize,因为会损失掉某些频率域里面的信息。

收获

1、要掌握合适的经验信息
2、把这些经验信息加入到损失函数中,对CNN进行指导

参考文献重点摘录可作为以后读

Residual image
[11] A. Pinto, W. Schwartz, H. Pedrini, and A. Rocha, “Using visual rhythms for detecting video-based facial spoof attacks,” IEEE Transactions on Information Forensics and Security, vol. 10, no. 5, pp. 1025–1038, may 2015.
[19] D. Cozzolino, D. Gragnaniello, and L. Verdoliva, “Image forgery detection through residual-based local descriptors and block-matching,” in IEEE International Conference on Image Processing, october 2014, pp. 5297–5301 Domain-information
[12] Y. Qian, J. Dong, W. Wang, and T. Tan, “Deep learning for steganalysis via convolutional neural networks,” in IS&T/SPIE Electronic Imaging, 2015, pp. 94 090J–94 090J.
[13] B. Bayar and M. Stamm, “A deep learning approach to
universal image manipulation detection using a new convolutional layer,” in ACM Workshop on Information Hiding
and Multimedia Security, 2016, pp. 5–10.

你可能感兴趣的:(anti-spoofing,论文阅读,antispoofing,论文)