MotiMotion: Motion-Controlled Video Generation with Visual Reasoning

TL;DR

MotiMotion introduces a reasoning-based framework for motion-controlled video generation that enhances plausibility and interaction realism by refining trajectories and hallucinating secondary motions.

Contribution

The paper presents a novel reasoning-then-generation approach, a confidence-aware control scheme, and a new benchmark for more realistic motion-controlled video synthesis.

Findings

Produces more plausible object behaviors and interactions.

Outperforms existing methods in human evaluations.

Demonstrates effectiveness on the new MotiBench dataset.

Abstract

Current motion-controlled image-to-video generation models rigidly follow user-provided trajectories that are often sparse, imprecise, and causally incomplete. Such reliance often yields unnatural or implausible outcomes, especially by missing secondary causal consequences. To address this, we introduce MotiMotion, a novel framework that reformulates motion control as a reasoning-then-generation problem. To encourage causally grounded and commonsense-consistent interactions, we leverage a training-free vision-language reasoner to refine image-space coordinates of primary trajectories and to hallucinate plausible secondary motions. To further improve motion naturalness, we propose a confidence-aware control scheme that modulates guidance strength, enabling the model to closely follow high-confidence plans while correcting artifacts under low-confidence inputs with its internal generative priors. To support systematic evaluation, we curate a new image-to-video benchmark, MotiBench, consisting of interaction-centric scenes where new events are triggered by motion. Both VLM-based evaluation and a human study on MotiBench demonstrate that MotiMotion produces videos with more plausible object behaviors and interaction, and is preferred over existing approaches.