hongwenzhang / ect-facealignment

[TIFS 2018] Combining Data-driven and Model-driven Methods for Robust Facial Landmark Detection

Combining Data-driven and Model-driven Methods for Robust Facial Landmark Detection

This is the demo code for the ECT approach described in Combining Data-driven and Model-driven Methods for Robust Facial Landmark Detection.

Requirements

• python 2.7

Instructions

You may need to compile the caffe firstly before you run the demo code. The pre-trained caffemodel could be downloaded from here.

cd caffe/python
for req in $(cat requirements.txt); do pip install $req; done
cd ..
make all
make pycaffe
cd ..
cd landmark_detection
python run_demo.py --imgDir path/to/you/testing/images --model path/to/the/pretrained/caffemodel --verbose True

Citation

If this work is helpful in your research, please cite the following paper.

@article{zhang2018combining,
  title={Combining data-driven and model-driven methods for robust facial landmark detection},
  author={Zhang, Hongwen and Li, Qi and Sun, Zhenan and Liu, Yunfan},
  journal={IEEE Transactions on Information Forensics and Security},
  volume={13},
  number={10},
  pages={2409--2422},
  year={2018},
  publisher={IEEE}
}

Acknowledgment

The code is developed upon Caffe-heatmap, Menpo, and Menpofit. Thanks to the original authors.

ECT was extended to detect facial landmarks on artistic portraits by Yaniv et al. Have a look at their code here.

ect-facealignment's Issues

The python __import__() show error. ImportError: No module named copy_reg.

Thanks for you sharing . I met one problem , when I ran the run_demo.py .

WARNING: Logging before InitGoogleLogging() is written to STDERR
W1026 14:04:18.403190 17798 _caffe.cpp:122] DEPRECATION WARNING - deprecated use of Python interface
W1026 14:04:18.403214 17798 _caffe.cpp:123] Use this instead (with the named "weights" parameter):
W1026 14:04:18.403218 17798 _caffe.cpp:125] Net('../caffe/models/300w/matlab.prototxt', 1, weights='../model_data/')
W1026 14:04:18.404358 17798 upgrade_proto.cpp:71] Note that future Caffe releases will only support input layers and not input fields.
Traceback (most recent call last):
File "run_demo.py", line 126, in
main(parser.parse_args())
File "run_demo.py", line 47, in main
fit_model = pickle.load(fit_model_file)
File "/usr/lib/python2.7/pickle.py", line 1384, in load
return Unpickler(file).load()
File "/usr/lib/python2.7/pickle.py", line 864, in load
dispatchkey
File "/usr/lib/python2.7/pickle.py", line 1096, in load_global
klass = self.find_class(module, name)
File "/usr/lib/python2.7/pickle.py", line 1130, in find_class
import(module)
ImportError: No module named copy_reg

能够获取每个点否是为遮挡点？

你好，谢谢你的开源。
我拜读了你的论文，每个点是否是遮挡点应该有概率输出的，这边如何提取这个信息？
给您添麻烦啦！
@HongwenZhang

python3 is ok?AttributeError: 'CLM' object has no attribute 'scales'

thanks for you sharing. I met one problem. AttributeError: 'CLM' object has no attribute 'scales'. Could you hepl me

运行run_demo.py得到的68个关键点的顺序，跟dlib检测的顺序一致吗？

跟这个顺序一样的吗？