DeepIncision

VLFM-Driven Efficient Recognition of Surgical Incisions

Environment This repo requires Pytorch>=1.9 and torchvision. We recommend using Docker to setup the environment. You can use this pre-built Docker image docker pull pengchuanzhang/maskrcnn:ubuntu18-py3.7-cuda10.2-pytorch1.9 or this one docker pull pengchuanzhang/pytorch:ubuntu20.04_torch1.9-cuda11.3-nccl2.9.9 depending on your GPU.

Then install the following packages:

pip install einops shapely timm yacs tensorboardX ftfy prettytable pymongo
pip install transformers 
python setup.py build develop --user

Backbone Checkpoints. Download the ImageNet pre-trained backbone checkpoints into the MODEL folder.

mkdir MODEL
wget https://penzhanwu2bbs.blob.core.windows.net/data/GLIPv1_Open/models/swin_tiny_patch4_window7_224.pth -O swin_tiny_patch4_window7_224.pth
wget https://penzhanwu2bbs.blob.core.windows.net/data/GLIPv1_Open/models/glip_tiny_model_o365_goldg_cc_sbu.pth

Command. train see

bash surgin_train.sh

val see

bash test.sh

metric

python metric.py

ensemble see

python metric_em.py

If you find this repo useful for your research, please consider citing our papers:

@article{zhaovision,
  title={Vision-language foundation model-driven efficient recognition and home-based management of surgical incisions},
  author={Zhao, Chunlin and Yi, Huahui and Jiang, Zekun and Yang, Mei and Yang, Yi and Guo, Yuchen and Wang, Jing and Yuan, Linyan and Chen, Xiao and Yang, Xue and others},
  journal={International Journal of Surgery},
  pages={10--1097},
  publisher={LWW}
}