Face Type Detector Using Active Shape Models with Stasm
Active Shape Model
Machine Learning
Installation
sudo apt-get install python-numpy
sudo apt-get install python-opencv
sudo pip install PyStasm
sudo pip install -U scikit-learn
##README Training data and model is not included because this is for proprietary application.
If you have some training data, you can make model & test what your face type is.
If you want to use this code in proprietary applications, please contact me
##Output
##Face types
- Oval
- Circle
- Almond
- Rectangle
- Squre
- Triangle/Diamond
##Trainig data format
-
/train_img
-
Include training images, indexing type is free, but same as included .txt in train_tag folder
-
Ex) tr_0001.jpg, tr_0002.jpg, ..., tr_xxxx.jpg
See /train_img/train0001.jpg
-
-
/train_tag
-
Include tag files, one to one correspondence between image and tag file.
-
Each tag file is include face type
-
Ex) tr_0001.txt, tr_0022.txt, ..., tr_xxxx.txt
See /train_tag/train0001.txt
-
##Usage:
Usage:
python main.py <command> [option] [path1] [path2]
Commands:
-h show help.
-T detect face line and color. need img file path(path1 = EXAMPPLE.jpg)
-Tv Same as -T, but show landmarks img
-tr training face line detector. need folder path(path1 = ./img, path2 = ./tag