av1_roi_lanczos_unsharp

[amoghmunikote]

submission name:

av1_roi_lanczos_unsharp

upload zipped archive.zip

archive.zip

report.txt

=== Evaluation results over 600 samples ===
  Average PoseNet Distortion: 0.07063574
  Average SegNet Distortion: 0.00508374
  Submission file size: 896,107 bytes
  Original uncompressed size: 37,545,489 bytes
  Compression Rate: 0.02386723
  Final score: 100*segnet_dist + √(10*posenet_dist) + 25*rate = 1.95

does your submission require gpu for evaluation (inflation)?

no

did you include the compression script? and want it to be merged?

yes

additional comments

SVT-AV1 (libsvtav1) with preset 0, CRF 33, film-grain synthesis 22, and ROI-aware preprocessing that denoises outside a hand-tuned driving corridor mask to save bits on irrelevant regions. Frames are downscaled to 45% with Lanczos before encoding. Decode uses Lanczos upscale back to original resolution with a 9-tap binomial unsharp mask at 40% strength to recover edge definition for SegNet.

[github-actions]

Thanks for the submission @amoghmunikote! 🤏

A maintainer will review your PR shortly.

To run the evaluation, a maintainer will trigger the eval workflow with your PR number.