junpan19 / rsdn

Code for paper "Video Super-Resolution with Recurrent Structure-Detail Network" . The code is under preparation and is coming soon.

Video Super-Resolution with Recurrent Structure-Detail Network (RSDN)

The official source code for the [Video Super-Resolution with Recurrent Structure-Detail Network] which is accepted by [ECCV-2020].

Train

We utilize 8 Nvidia Tesla V100 GPUs for training.

TODO

Test

We utilize 1 P100 GPU for testing. Test the trained model with best performance by

python test.py

rsdn's Issues

Help regarding inference

Can you provide a walkthrough on how to run inference on a custom test set?
Thanks.

Why is there a big gap in the results of training the network you provide yourself?

Why is there a big gap in the results of training the network you provide yourself?

LR_D = LR - LR_S RuntimeError: The size of tensor a (83) must match the size of tensor

File "/workspace/HOSTDIR/s_2021/RSDN-master/RSDN/load_test.py", line 58, in getitem
LR_D = LR - LR_S
RuntimeError: The size of tensor a (83) must match the size of tensor b (82) at non-singleton dimension 3

LICENSE

Hello,

and thank you for your work.
Could you please considering adding a license to your code?

Thanks,
D

How to calculate the parameters?

I run the code,it says "Model size: 23.57573M"

In the paper,We see Param about 3.83M/5.01M/6.19M

What is the conversion relationship between them？

Thank you!

when the code will be released?

code request

when will you release the code? can not wait to checkout!

when to realse the train script?

Hope to realse the code for train, i am so interested in your model.
when will the script release???????

bad results

My experimental results are quite different from those presented in the paper. The experimental results show that only the low-frequency information of the image is retained, but the high-frequency information is removed. I don't know if you have such a situation.

Your Results in New Super-Resolution Benchmarks

Hello,

MSU Graphics & Media Lab Video Group has recently launched two new Super-Resolution Benchmarks.

• Video Upscalers Benchmark: Quality Enhancement determines the best upscaling methods for increasing video resolution and improving visual quality.
• Super-Resolution for Video Compression benchmark aims to test Super-Resolution methods on compressed videos and select the best model for each video codec standard.

Your method achieved 14th place in Super-Resolution for Video Compression Benchmark in 'VVC compression' category. We look forward to your future work!

We would be grateful for your feedback on our work.

Image pre-processing

Hello,Why not call gaussian_downsample output boundary will appear a lot of ripples

Results in MSU Video Super Resolution Benchmark

Hello,
MSU Graphics & Media Lab Video Group has recently launched Video Super Resolution Benchmark.
We provide a comprehensive comparison of a variety of different VSR models.
You can see our current results: MSU VSR Benchmark.
If you are interested in participating, you can add your algorithm, following the submission steps.
You can read about MSU Graphics & Media Lab Video Group here and here.