Reproduce “Image Referenced Sketch Colorization Based on Animation Creation Workflow” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Dingkun Yan, Xinrui Wang, Zhuoru Li, Suguru Saito, Yusuke Iwasawa, Yutaka Matsuo, Jiaxian Guo
The authors present a diffusion method for colorizing animation sketches from image references while limiting interference from reference composition. It uses the sketch to guide drawing and separately processes masked foreground and background reference features with LoRA branches and a recovery transformer. On comparisons, it improves FID, MS-SSIM, and CLIP similarity over baselines, but not PSNR; results depend on mask quality.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.