search

guoyii / SACNN

Try to reproduce 2020_TMI_SACNN Self-Attention Convolutional Neural Network for Low-Dose CT Denoising with Self-supervised Perceptual Loss Network
MIT License
32 stars 5 forks source link

readme

V9 ===== `optim`：Adam `Percrtual loss`：False `d:p`：4：1 增加mse_loss * 100 Best: 195, 210, 235, 345<365> train in v6 V8 ===== `optim`：Adam `Percrtual loss`：True `d:p`：4：1 增加mse_loss * 50 Best: 255, 265, 270, 290, 300<305> train in v5 V7 ===== `optim`：Adam `Percrtual loss`：False `d:p`：4：1 增加mse_loss Best:365, 395, 420, 475, 495 V6 ===== `optim`：Adam `Percrtual loss`：True `d:p`：4：1 增加mse_loss Best: 315, 320*, 370, 400, 425 V5 ===== `optim`：Adam `Percrtual loss`：True `d:p`：4：1 Best:220, 250, 260, 265* 148.6753h V3 ===== `optim`：Adam `Percrtual loss`：False `d:p`：4：1 141.1220h Best:450, 445*, 320