ISTE

An Efficient Dual-branch Framework based on Implicit Self-texture Enhancement

[Paper][Code]

Dependencies

• numpy==1.23.4
• scikit_image==0.19.2
• torchvision==0.15.1
• torch==2.0.0
• pytorch-fid==0.2.1
• opencv-python==4.6.0.66

Train & Test

SwinIR-ISTE

Train: python train.py --config configs/train/train_swinir-iste.yaml --gpu 0,1,2,3

Test: python test.py --config configs/test/test-histo-2.yaml --model save/_train_swinir-iste/epoch-last.pth --window 0 --gpu 0

Demo For HistoSR at X2 Scale:

[Inference Demo] Verify Code: ctnv

Demo For TCGA at X2 Scale:

[Inference Demo] Verify Code: g9db