acne_detection
ran in python 3.6.8
- tensorflow == 1.13.1
export PYTHONPATH=$PYTHONPATH:/PATH_TO_THE_PROJECT/slim/
nohup python3.6 object_detection/model_main.py --pipeline_config_path=faster_rcnn_resnet101_coco.config --model_dir=./saved_models/ --num_train_steps=20000 --num_eval_steps=2000 --alsologtostderr > acne_train.log &python3.6 object_detection/export_inference_graph.py --input_type image_tensor --pipeline_config_path faster_rcnn_resnet101_coco.config --trained_checkpoint_prefix ./saved_models/model.ckpt-xxxxx --output_directory ./latest_models/A pretrained model (faster_rcnn_resnet101) can be found at MEGA, another (faster_rcnn_inception_v2) at MEGA.
P.S. MEGA is the best cloud drive I've ever used. Strong recommendation for it.
@article{thc_2022_acne_detection,
title = {{Acne Detection and Severity Evaluation with Interpretable Convolutional Neural Network Models}},
author = {Wen, Hao and Yu, Wenjian and Wu, Yuanqing and Zhao, Jun and Liu, Xiaolong and Kuang, Zhexiang and Fan, Rong},
journal = {Technology and Health Care},
doi = {10.3233/thc-228014},
issn = {1878-7401},
year = {2022},
month = {2},
publisher = {{IOS Press}},
volume = {30},
pages = {143--153}
}