This is the official repository for our work; The _ddd17 folder is used to process the ddd17 dataset. (1) Setup this code with pytorch=1.10.1,py3.8, cuda11.3, cudnn8.2.0_0 pip install apex,tensorBoard
(2) Usage
Training/Evaluation
Download the Camvid dataset http://mi.eng.cam.ac.uk/research/projects/VideoRec/CamVid/
bash multiscale_fullspike_seg/MIMO-UNet-Inverse.sh
--model_name "choose in the models dir"
--mode "train,val,test"
--exp_name "create new model save-file here"
--data_dir "dataset dir"
--resume "model.pkl"
--attention_map_dir "spike map dir"
(3)Visualize It can be downloaded from https://github.com/luo3300612/Visualizer Thanks for their nice contribution.