Comments (13)
REDS trained EDVR needs to be run through two stages, so you could try passing the output from stage 1 through the stage 2 model.
Other than that I believe there is a flip_test mode, which helps improve the quality.
from mmagic.
Yeah I figured out it has two stages afterwards, I'll rerun the experiment again soon.
The flip_test mode looks interesting, I guess an ensemble should increase generalisation a little.
from mmagic.
I can report a similar issue, I have tried using both stages, using flip_test, tested on a variety of videos. The model does not perform to the level it does on REDS4. It has multiple artifacts, and the quality overall is blurry.
from mmagic.
Okay, I solved my issue. The problem was the downsampling method. The datasets, that the model was trained on, were created by downsampling with MATLAB's imresize
function. So if you generate input data with anything else (opencv, ffmpeg) - it doesn't work. You have to use MATLAB's imresize
, or it's python's equivalent, which is implemented in this repo here.
from mmagic.
Hmm, that's what I feared. That kind of defeats the purpose of super resolution. I don't want to downsample my data, I want to upsample it :).
from mmagic.
@adamsvystun
Could you mention the exact flow that you used with the function you mentioned. did you basically send your input video through that method say if it H x W -> Target Resolution and subsequently pass it through EDVR to get the output. There were some weird blue-green artifacts during fast motion in my output, so I'm curious.
from mmagic.
@jorenvs It should work with upsampling. In my case I had a video in 720p and wanted to test 180p->720p upsampling, that's why I had to downsample. And it turns out that the model is very sensitive to the way you are doing this. If you have a video in low res only, it should just work.
@SreeHarshaNelaturu Yeah, for testing, I first downsample, then upsample with the model, and compare the results. Not sure about blue-green artifacts, I did not have any.
from mmagic.
well, my videos are 1344x1344, so not really low quality. That's all relative to the angle of the lens of course, these are generated from 360° 5.6k gopro videos. The goal is to be able read far away text on traffic signs and such.
from mmagic.
Thank you for the prompt response @adamsvystun I was wondering about the part you'd mentioned about not using FFmpeg or cv2 to generate input data. What did you use to extract frames from the video to SR in your case to using those methods.
from mmagic.
@SreeHarshaNelaturu I said don't use FFmpeg or cv2 for downscaling (resizing down). For frame extraction you can use anything you want.
from mmagic.
Gotcha, I think the blue-green error is a consequence of something else.
And yep, I was resizing via FFmpeg, might help to resize after extraction.
Thank you!
from mmagic.
I'm not sure but I think downscaling matlab's method could differ from FFmpeg and cv2
In my case, EDVR works well with the bicubic downscaling method but it has artifact like this with others (ex: low-res videos from youtube).
I guess that EDVR which trained with REDS dataset is overfitted about the reconstruction of bicubic downscaling as REDS dataset is consist of the bicubic downscaled dataset
from mmagic.
Yes, the current CNN-based methods does generalize to other datasets with different downsampling kernels.
There is another research filed called blind SR to solve this issue.
from mmagic.
Related Issues (20)
- Test/Inference without metrics>s HOT 4
- [Bug] How to set random_seed HOT 2
- [Feature] how to use gradio_inpainting.py? HOT 2
- [Bug] Controlnet weights are saved as 'controlnet.module.model...' but are indeed loaded as 'controlnet.model...'. HOT 2
- sr的save_image色彩不对[Bug] HOT 1
- The color of the saved image of the super-resolution model is wrong[Bug] HOT 1
- mim install 'mmcv>=2.0.0' time out
- [Feature] Difference between inputs and data_samples HOT 1
- How to set that not save the picture of the verification result during the training process[Docs] HOT 5
- [Docs] Please update TUTORIAL 3: INFERENCE WITH PRE-TRAINED MODELS of the docs HOT 2
- [Bug] When using resmue to continue training, AssertionError: If capturable=False, state_steps should not be CUDA tensors. occurs HOT 1
- How to use y channel training[Bug] HOT 3
- [Bug] raise ValueError(f'Unknown inferencer task: {self.task}') ValueError: Unknown inferencer task: Text2Image, Inpainting HOT 3
- [Bug] HOT 20
- Can video hyperprocessors, such as realbasicvsr, only use sequence images instead of video when testing? In addition, is there a size limit for the image as input? Can the image of super size (such as 10000*10000) be directly used as input? If you can not directly as input, then how to modify the code to achieve large image input[Feature]
- super-large images(10000*10000) problem
- [Bug][Controlnet Animation]control_detector in anythingv3_config.py should be 'lllyasviel/Annotators'
- I followed the official installation tutorial and tried 3 times to keep this error HOT 2
- RTX3090 run video spuer resulotion inference error HOT 2
- [Bug]
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from mmagic.