Deep Compressive Sensing for Visual Privacy Protection in FlatCam Imaging
Deep Compressive Sensing for Visual Privacy Protection in FlatCam Imaging
复制标题
DOI：
10.1109/iccvw.2019.00492
复制
发表时间：
2019-10
期刊：
影响因子：
--
通讯作者：
Thuong Nguyen Canh;H. Nagahara
中科院分区：
文献类型：
--
作者：
Thuong Nguyen Canh;H. Nagahara
在传统的隐私相机中，检测之后的投影很容易受到软件攻击，威胁到图像传感器数据的暴露。通过用编码掩模复用入射光，FlatCam相机消除了空间相关性，并捕获了视觉上受保护的图像。然而，FlatCam成像存在重构质量差、不注重视觉信息隐私等问题。在本文中，我们提出了一种基于深度学习的压缩感知方法来重建和保护敏感区域免受安全FlatCam测量的影响。我们通过面部分割预测敏感区域，并将其从捕获的测量中分离出来。利用模拟数据对深度压缩感知网络进行了训练，并在模拟和真实的FlatCam数据上进行了测试。
Detection followed by projection in conventional privacy cameras is vulnerable to software attacks that threaten to expose image sensor data. By multiplexing the incoming light with a coded mask, a FlatCam camera removes the spatial correlation and captures visually protected images. However, FlatCam imaging suffers from poor reconstruction quality and pays no attention to the privacy of visual information. In this paper, we propose a deep learning-based compressive sensing approach to reconstruct and protect sensitive regions from secured FlatCam measurements. We predict sensitive regions via facial segmentation and separate them from the captured measurements. Our deep compressive sensing network was trained with simulated data, and was tested on both simulated and real FlatCam data.