FramePrompt: In-context Controllable Animation with Zero Structural Changes

TL;DR

FramePrompt is a minimalist framework that leverages pre-trained video diffusion transformers to generate controllable character animations from reference images and motion cues without structural modifications.

Contribution

It introduces a sequence-level visual conditioning approach that simplifies controllable animation by avoiding complex architectures and guider modules.

Findings

Outperforms baseline methods on multiple metrics

Simplifies training process

Effectively uses pre-trained models for animation control

Abstract

Generating controllable character animation from a reference image and motion guidance remains a challenging task due to the inherent difficulty of injecting appearance and motion cues into video diffusion models. Prior works often rely on complex architectures, explicit guider modules, or multi-stage processing pipelines, which increase structural overhead and hinder deployment. Inspired by the strong visual context modeling capacity of pre-trained video diffusion transformers, we propose FramePrompt, a minimalist yet powerful framework that treats reference images, skeleton-guided motion, and target video clips as a unified visual sequence. By reformulating animation as a conditional future prediction task, we bypass the need for guider networks and structural modifications. Experiments demonstrate that our method significantly outperforms representative baselines across various evaluation metrics while also simplifying training. Our findings highlight the effectiveness of sequence-level visual conditioning and demonstrate the potential of pre-trained models for controllable animation without architectural changes.