Block-wise confidence-aware CNN that predicts the compression quality level (e.g. JPEG quality factor) of an image in real time — official TensorFlow implementation of the BMVC 2021 paper
-
Updated
Jul 8, 2026 - Python
Block-wise confidence-aware CNN that predicts the compression quality level (e.g. JPEG quality factor) of an image in real time — official TensorFlow implementation of the BMVC 2021 paper
ANSI motion lab // lo-fi frames // cursed little files
PyTorch models for AV1 artifact removal. Features a universal CRF-Preset conditional model (FiLM-based U-Net) and ultra-light Nano models (U-Net, ResNet, ...) for fast real-time video restoration (different model for each CRF bucket).
AI Old Video Restorer for Windows focused on offline VHS and archive-video repair: denoise, deblur, color cleanup, face recovery, upscale presets, before/after preview and export profiles.
Stop the "Homeopathic 2K; 4K; 8K" Scam.
AI Old Video Restorer for Windows focused on offline VHS and archive-video repair: denoise, deblur, color cleanup, face recovery, upscale presets, before/after preview and export profiles.
To associate your repository with the compression-artifacts topic, visit your repo's landing page and select "manage topics."