# 3DMM-fitting
This project is designed to fit a 3DMM to a frontal-face picture profile and a profile picture of the same person simultaneously. This is supposed to lead to a more reliable fitting result than the traditional way in which only one 2D picture is used, since we acquired additional depth information from the extra profile image.
## Library requirements
* Python3.6
* [OpenCV](http://opencv.org/)
* [Dlib](http://dlib.net/)
* [Numpy](http://www.numpy.org/)
* [toml](https://github.com/uiri/toml)
The code has only been tested on Windows10 with Anaconda Python.
## Instructions
You can find sample code at `test\fitting_test.py`. In the folder `data\`, two sets of sample images are already given to test the code. These images are from [color FERET database](https://www.nist.gov/itl/iad/image-group/color-feret-database). The facial landmarks are saved as pts files with the same name as the pictures. Please note that the frontal-face landmarks are annotated according to the iBug but the profile landmarks are annotated in a new way showed as below.
![the landmarks of a profile](https://i.imgur.com/ARFkW5F.jpg)
Not all the landmarks are used in the process of 3D-fitting.
The frontal face image is automatically annotated with Dlib library. You can call the `frontal_face_marker` funtion at `assist\marker.py` to get a pts file contains the landmarks of the frontal face image. The profile image is presently marked manually. You can call the `manual_marker` fuction at `assist\marker.py` to do it.
## Presentation
Run `test\fitting_test.py` with default imput images, you should get a picture discribes the accuracy of the fitting.
![fitting result img](https://github.com/Yinghao-Li/3DMM-fitting/blob/master/test/00029ba010_960521-outcome.jpg
)
This picture will be saved in the `test\` folder, along with the generated 3D model as ply file.
![fitting result 3D](https://github.com/Yinghao-Li/3DMM-fitting/blob/master/test/3D-captured.PNG)
没有合适的资源?快使用搜索试试~ 我知道了~
温馨提示
3DMM模型是三维人脸重建的基础模型,用于三维人脸重建任务。(3dmm model is the basic model of 3D face reconstruction, which is used for 3D face reconstruction task.)
资源推荐
资源详情
资源评论
收起资源包目录
3DMM-fitting.zip (50个子文件)
3DMM-fitting
core
linear_shape_fitting.py 13KB
Blendshape.py 2KB
orthographic_camera_estimation_linear.py 4KB
LandmarkMapper.py 2KB
MorphableModel.py 3KB
Landmark.py 2KB
utils.py 3KB
Mesh.py 2KB
closest_edge_fitting.py 15KB
glm.py 5KB
__init__.py 0B
contour_correspondence.py 12KB
blendshape_fitting.py 7KB
RenderingParameters.py 7KB
PcaModel.py 5KB
fitting.py 55KB
EdgeTopology.py 2KB
render.py 37KB
data
00019pr010_940128.tif 96KB
00029ba010_960521.tif 96KB
00029pr010_940128.pts 476B
00019pr010_940128.pts 476B
00019fa010_940128.tif 96KB
00029ba010_960521.pts 1KB
00029pr010_940128.tif 96KB
00019fa010_940128.pts 1KB
test
draw_mesh_test.py 6KB
3D-captured.PNG 421KB
00029ba010_960521-output.ply 272KB
Display.ipynb 7KB
fitting_test.py 4KB
00029ba010_960521-outcome.jpg 257KB
py_share
py_expression_blendshapes_3448.bin 243KB
sfm_3448_edge_topology.json 1.78MB
adj_dict_3448.pkl 322KB
profile_to_sfm.txt 426B
sfm_model_contours.json 678B
ibug_to_sfm.txt 3KB
sfm_reference_symmetry.txt 3KB
py_sfm_3448_edge_topology.json 1.39MB
py_sfm_shape_3448.bin 7.65MB
assist
content_search.py 1KB
get_adj_dict.py 810B
__init__.py 0B
marker.py 10KB
model_transfer.py 2KB
search.ipynb 1KB
README.md 2KB
.gitattributes 66B
毕设论文-李英昊.pdf 3.01MB
共 50 条
- 1
资源评论
- Maisie·FCodeLab2022-12-14资源内容总结地很全面,值得借鉴,对我来说很有用,解决了我的燃眉之急。
wouderw
- 粉丝: 268
- 资源: 2960
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
最新资源
- :基于java打造的深度学习框架,帮助你快速搭建神经网络,实现模型推理与训练,引擎支持自动求导,多线程与GPU运算
- 第九次作业(XY图,XY图显示,三维曲面,数字波形图)
- 微信小程序实战案例:打造高效便捷的在线书店.zip
- 1.0.5win(1)(1).exe
- ESP8266 WiFi模块入门教程:从连接到配置.zip
- 词频统计:从基础到实践的应用指南.zip
- 滑动窗口:深入理解与应用.zip
- WRF(Weather Research and Forecasting Model)模型:深入解析与实用操作指南.zip
- IngeekDK.json
- DHT11温湿度传感器:原理、应用与操作指南.zip
资源上传下载、课程学习等过程中有任何疑问或建议,欢迎提出宝贵意见哦~我们会及时处理!
点击此处反馈
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功