VLight

Requirements

• Python 3.6 or newer
• PyTorch >= 1.1.0
• CUDA >= 10.0
• cuDNN >= 7.5

Getting started

Download or clone the full repository.

Install dependencies

pip install -r requirements.txt

Datasets

You can train models or run evaluations on the following vessel segmenation datasets:

• DRIVE
• CHASE_DB1
• HRF

Download the original datasets and place them in ./data. Alternatively, you can also update the paths in local_config.ini.

Evaluation

To evaluate the pre-trained models:

python eval_vessels.py --model vlight --modelname models/pretrained/{drive,chase,hrf}_vlight --dataset {drive,chase,hrf}

Training

To train VLight on DRIVE for 40 epochs (80k samples) with a learning rate of 0.001, for example, run:

python train_vessels.py --sessionname drive_vlight --lr 0.001 -e 40 --dataset-train drive --dataset-val drive