microsoft / microsoft/DCVC

The saved pictures pf DCVC-FM have obvious color differences.

Open
#85 6 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
829
Forks
138
Avg merge
1d 2h
Merged PRs (30d)
1

Description

I try to save the decoded frame of DCVC-FM of run_one_point_fast, so I add the code below:
recon_writer = PNGWriter(args['bin_folder'], args['src_width'], args['src_height'])
and

if save_decoded_frame:
                yuv_rec = x_hat.squeeze(0).cpu().numpy()
                rgb = ycbcr444_to_rgb(yuv_rec[:1, :, :], yuv_rec[1:, :, :])
                recon_writer.write_one_frame(rgb=rgb, src_format='rgb')

The metrics of saved json is as the paper report, but the result have obvious color differences and a 8dB PSNR drop.

Image

Contributor guide

Open the contributing guide

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

Start at the run_one_point_fast path and inspect the save_decoded_frame block using PNGWriter and ycbcr444_to_rgb. Compare the saved frame with the decoded tensor and the reported metrics, then verify that the saved image has matching colors and no PSNR drop.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
computer-vision, machine-learning
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
52/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.