我们使用视差效果在单个2D图像上提供3D体验，即用户可以移动其实时跟踪的面部以可视化深度效果。 通过使用修复方法，该程序可以绘制丢失的背景，该背景现在可以通过用户的移动来访问。
修复方法和大部分代码取自：
Shih3DP20, Shih, Meng-Li and Su, Shih-Yang and Kopf, Johannes and Huang, Jia-Bin, 3D Photography using Context-aware Layered Depth Inpainting, IE
ImageNet中人脸混淆的研究
论文代码：
，，，和
article{yang2021imagenetfaces,
title={A Study of Face Obfuscation in ImageNet},
author={Yang, Kaiyu and Yau, Jacqueline and Fei-Fei, Li and Deng, Jia and Russakovsky, Olga},
journal={arXiv preprint arXiv:2103.06191},
Led3D
This project is an implementation for "Led3D: An lightweight and efficent deep approach to recognizing low-quality 3d faces" [ ], which is accepted by CVPR2019.
Dataset:
<<<<<<< HEAD
Function
3D Face Preprocessing (Done.
