yanchaoz / CFDA-M

CFDA-M

Codes and data for "CFDA-M: Coarse-to-Fine Domain Adaptation for Mitochondria Segmentation via Patch-wise Image Alignment and Online Self-training" (2022 IEEE International Conference on Bioinformatics and Biomedicine)

Datasets

Data used in this paper can be download here.

Dependencies

Training

Coarse stage

cd /stage1

CUDA_VISIBLE_DEVICES=0 python train.py \
--name experiment_cut2seg \
--raw_A_dir ./preprocess/VNC2Lucchi/VNC/ \
--raw_A_seg_dir ./preprocess/VNC2Lucchi/VNC/ \
--raw_B_dir ./preprocess/VNC2Lucchi/Lucchi/ \
--sub_list_A ./preprocess/VNC2Lucchi/train_VNC.txt \
--sub_list_B ./preprocess/VNC2Lucchi/train_Lucchi.txt \
--batch_size 4 \
--angle 0 \
--model cut2seg_model_train \
--netG resnet_9blocks \
--netD basic \
--netS duseunet \
--pool_size 50 \
--no_dropout \
--dataset_mode cut2seg_train \
--input_nc 1  \
--output_nc 1 \
--output_nc_seg 1 \
--lambda_GAN 1.0 \
--lambda_NCE 1.0 \
--lambda_DICE 1.0 \
--lambda_SC 1.0 \
--checkpoints_dir ./checkpoints/VNC2Lucchi/ \
--display_id 0

Fine stage

cd /stage2

CUDA_VISIBLE_DEVICES=0 python train.py

Acknowledgement

This code is based on AccSeg-Net (MICCAI'21) by Bo Zhou et al. Should you have any further questions, please let us know. Thanks again for your interest.