OpenImagingLab / OpenImagingLab/FlashVSR

Distortion in face and text under 1.5x upscale

Open
#40 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
1.9k
Forks
152
PR merge metrics
No merged PRs in 30d

Description

Hi, thanks for your great work! I’m using the model for video super-resolution with a fixed 1.5× upscale ratio(480p -> 720p), which I can’t change due to task requirements. The overall results are good, but local facial expressions and text sometimes appear distorted (similar to the issue in #25). I’ve already experimented with several hyperparameters (such as sparse ratio and local range), but the results didn’t show much difference. Are there any other settings or parameters I could adjust to improve detail fidelity in this case?

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No file, test, or entry point is identified in the issue. Reproduce the fixed 1.5× 480p-to-720p case, compare the facial and text distortions with issue #25, and review the reported sparse-ratio and local-range experiments. Done means identifying a supported setting that improves detail fidelity or documenting that the limitation remains.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
computer-vision, machine-learning
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.