Face2to3D

KO | EN

gitlite — search

#python #ai #llm #golang #android #rust #linux #macos #nodejs #java #cli #kubernetes #security #openai

(★ 14)

3D Facial Reconstruction

Face2to3D Latest Version Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Face2to3D
3D Facial Reconstruction
资源及相关链接放在[material](https://github.com/changebio/Face2to3D/tree/master/material)文件夹

代码放在[sources](https://github.com/changebio/Face2to3D/tree/master/sources)文件夹

自己代码可以在[try](https://github.com/changebio/Face2to3D/tree/master/try)文件夹中以自己的名字命名一个文件夹

## 重要信息
jupyter notebook address

http://jupyter-uaitrain-bj2.ucloud.cn:82/ff25f3e5-1b23-43df-b24d-82f963c81ba5/?token=54d44cc6eceb97ba9c6f9ad37beb675273df8ad937c54f19

```
#Jupyter workflow in remote server after login jupyter
data -> Face2to3D -> try ->your_code_directory
#Terminal workflow in remote server
New -> Terminal
cd /data/data/Face2to3D/try

#conda env in server (base, caffe2, mxnet, tf1.9)
conda activate base # for tensorflow2.0, keras
conda activate caffe2 # for Pytorch, caffe2
conda activate mxnet # for mxnet
conda activate tf1.9 # for tf1.9

```
##### [How can I use version control to collaborate with other people?](https://swcarpentry.github.io/git-novice/08-collab/index.html)

#### 注意事项
1.尽量不要把**代码或文件**放到data文件夹外。因为可能**丢失文件**。

2.在**try**文件夹中的子项目文件夹写和修改代码。尽量不要在sources中修改代码。

3.合作项目代码和资源尽量更新到GitHub，避免大家做重复工作。

**4.记着自己打开的Jupyter Notebook和terminal，避免误会其他人的。（Jupyter Notebook命名规则：名字简写_其他，如 hy_3ddfa.ipynb**

## State-of-Art Models Code
### [2DASL](https://github.com/changebio/Face2to3D/tree/master/sources/2DASL)

### [3DDFA](https://github.com/changebio/Face2to3D/tree/master/sources/3DDFA)
[Offical Website with Code and Data](http://www.cbsr.ia.ac.cn/users/xiangyuzhu/projects/3DDFA/main.htm)

### [PRNet](https://github.com/changebio/Face2to3D/tree/master/sources/PRNet) 
[pytorch version](https://github.com/changebio/Face2to3D/tree/master/sources/pytorch-prnet) 
[TF training code](https://github.com/changebio/Face2to3D/tree/master/sources/training_codes_for_PRNet_3D_Face)

### [Nonlinear 3DMM](https://github.com/changebio/Face2to3D/tree/master/sources/Nonlinear_Face_3DMM)

### [VRN](https://github.com/changebio/Face2to3D/tree/master/sources/vrn)

## 项目说明
吾日三省吾身：白否？富否？美否？ 我们将利用三维人脸重建技术，更好地“审视”自己。三维人脸重建是3D计算机视觉中的重要部分，可以广泛应用于三维人脸识别、辅助医疗、个性化3D打印、影视特效、虚拟现实等领域。传统三维人脸重建技术需要根据多视图估计相机运动，并计算每张图像的深度图，需要消耗大量时间及计算资源。利用深度摄像头可以更快速获取图像的深度，在一定程度上减少重建时间，然而对于普通摄像头而言，如何快速重建出三维人脸模型？近年来，随着深度学习技术的发展，从视频中快速重建出三维人脸模型已经变得可能。看腻了二维人脸图像，不妨来挑战一下更高级的3D人脸建模。 
本课题涉及：人脸检测；人脸特征点定位；亚洲人脸模型构建；人脸重建。

![项目导图](https://github.com/changebio/Face2to3D/blob/master/material/IMG_2956.JPG)

Latest Network Architecture and Performance Comparison
![Performance](https://user-images.githubusercontent.com/8948023/56006880-36622000-5d09-11e9-9465-8d52e3433d5f.png)

## 数据

**3D Morphable Model (3DMM)**

- Basel Face Model (BFM)
	- BFM2009 (53490vertices) [网盘](https://pan.baidu.com/s/1DHzZUmut-kOSC6AZGWrG1g) n4we 
	- BFM2017 (53149vertices)

- Surrey Face Model (SFM)
	- SFM1724
	- SFM3448
	- SFM16759 (No texture) 
	- SFM29587 (No texture)

- Large scale Face Model (LSFM)

## 计划

## 小组

## 建议

## Tools
### common git command
```
git status
git add .
git commit -m "xxxxxx"

#pull and push code from GitHub
git pull #equal to git fetch; git merge
git push
```

### basic markdown format

https://help.github.com/en/articles/basic-writing-and-formatting-syntax

Face2to3D

Face2to3D Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Face2to3D

Face2to3D Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Download README