How to test my own videos, do I need to train first?
I only have a low res video

Hello there,

I'm a begginer programmer trying to reproduce your result. However, i met some difficulties wich i think comme from the dataset.

here is my branche organisation after extracting the .zip :

RefVSR
├── evaluation
├──install
├──ckpt
├──...
├──DATA_OFFSET
├── source
├──target
├──RealMCVSR.z01
├── ...
├──RealMCVSR.z21

i have finally run this scipt : /scripts_eval/eval_RefVSR_MFID_8K.sh with the following configuration

And here is the error message :

Could you help or give me another direction to resolve my bug ?

your faithfully,

Maxx

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.

If you are interested in participating, you can add your algorithm following the submission steps:

• Submit for Video Upscalers Benchmark: Quality Enhancement
• Submit for Super-Resolution for Video Compression benchmark

We would be grateful for your feedback on our work!

I have downloaded the pre-trained models as well as dataset from the given links and tried to run the evaluation scripts( I didn't modify any hyperparameters except for the dataset path and log path ). However, there's a large gap between my evaluation results and those in the paper.

So I would like to ask what is the problem and what should I try to get the results in the paper?

Thank you!

Hello, do low resolution images and reference images have to be the same size in inferenceing? If my reference image is not the same size as the low resolution image, how can I inference?

Hi, have you released the iOS app you made to concurrently record videos of three cameras on iPhone 12 Pro Max?

Welcome update to OpenMMLab 2.0

I am Vansin, the technical operator of OpenMMLab. In September of last year, we announced the release of OpenMMLab 2.0 at the World Artificial Intelligence Conference in Shanghai. We invite you to upgrade your algorithm library to OpenMMLab 2.0 using MMEngine, which can be used for both research and commercial purposes. If you have any questions, please feel free to join us on the OpenMMLab Discord at https://discord.gg/A9dCpjHPfE or add me on WeChat (ID: van-sin) and I will invite you to the OpenMMLab WeChat group.

Here are the OpenMMLab 2.0 repos branches:

Attention: please create a new virtual environment for OpenMMLab 2.0.

Thank you for your great work.
I'm going to reproduce results in Table 2, but I'm confused in some configurations in other models.

1. How did you configure the frame_num in other benchmark models? I guess this hyper-parameter would be important, but there is no information about this in the paper.
2. Why the frame_num of each model is different? It varies from 7 to 13.

Thank you

Do I just need to download a RealMCVSR.zip file, or do I need to download all five files?

Thanks for your great work. From your results in Table 2, it seems that the model using l1 loss (Ours-l1) could outperform the model using the proposed two-stage training strategy (Ours) over 3 dB, and it seems an one-stage training process from your training code.

So,

1. Why does the model “Ours-l1” perform better than the model “Ours”? It seems that you don't have the groundtruth of real-world HR_UW.

2. How does one-stage training process works?

I see it was tested under Ubuntu. But is it possible to run on Windows?

Hi, I download the dataset, but they both fail in the middle stage and prompt unknown server error. Could you please check the download link and ensure they are properly downloaded.

First of all, thanks for your great work. Your paper was intereseting, and results were great!
I was trying to use your code, especially datasets.py and get_patch method, but faced one problem.

In the train time (cropped)

• LR_UW size: (64, 64)
• LR_REF_W size: (128, 128)

In the test time

• LR_UW size: (480, 270)
• LR_REF_W size: (480, 270)

I understand that it is because of the cropping done in get_patch. For the W reference images, I found that your code gets twice a larger patch than UW images. However, my concerns is that why ratio of reference image and LR image is different during train time and test time. More precisely,

1. Is the ratio of reference image and LR image are intended to be different during train and test time?
2. Then how do your model handle such different ratio?
3. If not intended, which is right? Or is there anything I missed?

I'm using your default config, and flag_HD_in is false. Thank you :)