Face Detect Landmarks 200 Points

About

本项目使用dlib人脸特征识别，在原来的68点的基础上扩展到200点

Features

• 200个特征点，覆盖全脸、眼部、眉毛、嘴及嘴唇、鼻子包含轮廓及鼻梁骨架

Getting Start

1. 下载illinois的图片
2. 安装必要的插件,如dlib, opencv
3. 运行python脚本test_shape_predictor.py

在根目录已经提供了一个用如下参数训练出来的预测模型，也可以自己通过train_shape_predictor.py来训练

    options.nu = 0.05
    options.tree_depth = 2
    options.be_verbose = True

Demo

Contact

• Watson Wuh

---

#face detect landmarks 200 points

About

base on dlib's face landmarks detecting from 68 exntend to 200 points

Getting start

Actually, in this project already provide predictor model trained with

    options.nu = 0.05
    options.tree_depth = 2
    options.be_verbose = True

plese follow below steps:

1. download images at Helen dataset from illinois
2. install required plugins, like dlib, opencv
3. run the python script test_shape_predictor.py

Actually, in this project already provide predictor model trained by train_shape_predictor.py with

    options.nu = 0.05
    options.tree_depth = 2
    options.be_verbose = True

Demo

Contact

• Watson Wuh