VISTA: Triplet-Supervised Video Style Transfer with Diffusion Transformers

TL;DR

VISTA introduces a large synthetic dataset and a diffusion-transformer framework for improved, consistent video style transfer that effectively disentangles style, content, and motion.

Contribution

The paper presents VISTA-1000, a synthetic triplet dataset, and a novel diffusion-transformer model for robust, temporally consistent video style transfer.

Findings

Achieves state-of-the-art style fidelity and temporal consistency.

Outperforms existing methods in content preservation.

Robust under occlusions and complex motions.

Abstract

Video style transfer aims to render videos in a target artistic style while preserving content, structure, and motion. While image stylization has advanced rapidly, video stylization remains challenging due to temporal inconsistency. Most existing methods stylize frames or keyframes and enforce consistency via heuristic temporal propagation, which is brittle under occlusions, disocclusions, and long-term motion, leading to drift and flickering artifacts. We argue that a fundamental bottleneck lies in the lack of large-scale triplet data and a principled training paradigm that jointly models and disentangles style, content, and motion.To address this, we introduce VISTA-1000, a synthetic dataset with 1,000 styles and motion-aligned triplets of style reference, clean video, and stylized video, and propose a diffusion-transformer-based in-context video style transfer framework with a lightweight style adapter for robust style extraction. Extensive experiments demonstrate SOTA performance in style fidelity, temporal consistency, and content preservation.