PeCA: Palette Context Assisted Inference for Test-Time Paint-Bucket Colourisation on Animation Videos

2026-08-01

AI Summary

This work addresses the instability in inter-frame color correspondence commonly encountered in automatic coloring of hand-drawn animation, which often arises from region fragmentation and lack of contextual information. The authors propose a training-free, plug-and-play inference framework that, for the first time, introduces a palette-based contextual reasoning mechanism during test time. By integrating spatio-temporal context with semantic palette information, the method enables robust color propagation and assignment across frames. It achieves substantial improvements in both accuracy and temporal consistency for long-form animated sequences, demonstrating consistent performance gains on existing benchmarks as well as a newly introduced long-video test set, thereby validating its effectiveness and generalization capability.

Abstract

In animation production, paint-bucket colourisation for hand-drawn animation is a labour-intensive procedure that assigns each enclosed region in line sketches a colour from reference design sheets. Recent automatic paint-bucket colourisation pipelines mirror this workflow via region correspondence, but correspondences can be brittle when regions are ambiguous fragments without proper context. In this paper, we propose Palette Context Assisted (PeCA), a new training-free, plug-and-play framework for animation video colourisation that aims to close this gap at test-time via reasoning over spatial and temporal contexts. Extensive experiments on existing benchmarks and a newly introduced long-video test case show consistent performance boosts.

Problem

• paint-bucket colourisation
• animation videos
• region correspondence
• context ambiguity
• test-time colourisation

Innovation

• test-time colourisation
• palette context
• animation video
• region correspondence
• training-free framework