jonzhaocn / fbpconvnet_pytorch

FBPConvNet pytorch

A pytorch implement of FBPConvNet.

Dataset

dataset: https://drive.google.com/open?id=0B9fSVcoxJuVwMDlxbXdvcTRaM2M [from panakino/FBPConvNet ]

• note : phantom data set (x20) is only provided.

Examples

Noisy	Ground Truth	FBPConvNet Output
		SNR:22.063571
		SNR:22.175497
		SNR:24.397142
		SNR:25.136485

References

1. Jin K H, McCann M T, Froustey E, et al. Deep convolutional neural network for inverse problems in imaging[J]. IEEE Transactions on Image Processing, 2017, 26(9): 4509-4522.
2. https://github.com/panakino/FBPConvNet
3. https://github.com/panakino/fbpconv_tf

fbpconvnet_pytorch's Issues

The link to the dataset is unavailable

Hello, the link to the dataset you provided seems to be unavailable. Perhaps you could share a new link for downloading the dataset?

Please provide the training result.

I have tried the train.py. But after 97 epoch of training, the result is as the pic below:

and the training loss is as:

loss (epoch-93-iteration-300) : 13.986238
save image: ./samples/epoch-93-iteration-400.jpg
loss (epoch-93-iteration-400) : 1.800856
save image: ./samples/epoch-94-iteration-100.jpg
loss (epoch-94-iteration-100) : 7.462033
save image: ./samples/epoch-94-iteration-200.jpg
loss (epoch-94-iteration-200) : 1.865844
save image: ./samples/epoch-94-iteration-300.jpg
loss (epoch-94-iteration-300) : 2.170855
save image: ./samples/epoch-94-iteration-400.jpg
loss (epoch-94-iteration-400) : 3.993443
save image: ./samples/epoch-95-iteration-100.jpg
loss (epoch-95-iteration-100) : 15.650680
save image: ./samples/epoch-95-iteration-200.jpg
loss (epoch-95-iteration-200) : 1.744222
save image: ./samples/epoch-95-iteration-300.jpg
loss (epoch-95-iteration-300) : 1.496086
save image: ./samples/epoch-95-iteration-400.jpg
loss (epoch-95-iteration-400) : 1.489223
save image: ./samples/epoch-96-iteration-100.jpg
loss (epoch-96-iteration-100) : 2.223736
save image: ./samples/epoch-96-iteration-200.jpg
loss (epoch-96-iteration-200) : 1.954349
save image: ./samples/epoch-96-iteration-300.jpg
loss (epoch-96-iteration-300) : 2.035094
save image: ./samples/epoch-96-iteration-400.jpg
loss (epoch-96-iteration-400) : 26.680279
save image: ./samples/epoch-97-iteration-100.jpg
loss (epoch-97-iteration-100) : 5.275980
save image: ./samples/epoch-97-iteration-200.jpg
loss (epoch-97-iteration-200) : 9.626511

I do not know if everything is going well. So please provide the training results for reference.