Install dependencies for python 3.7 using:
Install dependencies for python 3.10 using:
You may need to install a different pytorch build, depending on your GPU support in CUDA https://pytorch.org/get-started/previous-versions/
# the dataset directory:
|-- ${image_dir}
|-- WFLW
| -- WFLW_images
|-- 300W
| -- afw
| -- helen
| -- ibug
| -- lfpw
|-- COFW
| -- train
| -- test
|-- ${annot_dir}
|-- WFLW
|-- train.tsv, test.tsv
|-- 300W
|-- train.tsv, test.tsv
|--COFW
|-- train.tsv, test.tsv
Dataset | Model |
---|---|
WFLW | google / baidu |
300W | google / baidu |
COFW | google / baidu |
python main.py --mode=train --device_ids=0,1,2,3 \
--image_dir=${image_dir} --annot_dir=${annot_dir} \
--data_definition={WFLW, 300W, COFW} \
--ckpt_dir=${out_dir}
The batch_size parameter may need to be set depending on available GPU memory. e.g "--batch_size=16"
python main.py --mode=test --device_ids=0 \
--image_dir=${image_dir} --annot_dir=${annot_dir} \
--data_definition={WFLW, 300W, COFW} \
--pretrained_weight=${model_path} \
--ckpt_dir=${out_dir}
python evaluate.py --device_ids=0 \
--model_path=${model_path} --metadata_path=${metadata_path} \
--image_dir=${image_dir} --data_definition={WFLW, 300W, COFW} \
--ckpt_dir=${out_dir}
To test on your own image, the following code could be considered:
python demo.py
The models trained by STAR Loss achieved SOTA performance in all of COFW, 300W and WFLW datasets.
Please consider citing our papers in your publications if the project helps your research. BibTeX reference is as follows.
@inproceedings{Zhou_2023_CVPR,
author = {Zhou, Zhenglin and Li, Huaxia and Liu, Hong and Wang, Nanyang and Yu, Gang and Ji, Rongrong},
title = {STAR Loss: Reducing Semantic Ambiguity in Facial Landmark Detection},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2023},
pages = {15475-15484}
}
This repository is built on top of ADNet. Thanks for this strong baseline.