VideoJAM: Joint Appearance-Motion Representations for Enhanced Motion Generation in Video Models

TL;DR

VideoJAM introduces a joint appearance-motion framework that improves motion coherence and visual quality in video generation by learning a combined representation and guiding generation with dynamic motion predictions.

Contribution

It presents a novel framework that integrates appearance and motion learning, with a new inference guidance mechanism, applicable to any video model without data or architecture modifications.

Findings

Achieves state-of-the-art motion coherence in video generation.

Enhances visual quality of generated videos.

Applicable to various models with minimal changes.

Abstract

Despite tremendous recent progress, generative video models still struggle to capture real-world motion, dynamics, and physics. We show that this limitation arises from the conventional pixel reconstruction objective, which biases models toward appearance fidelity at the expense of motion coherence. To address this, we introduce VideoJAM, a novel framework that instills an effective motion prior to video generators, by encouraging the model to learn a joint appearance-motion representation. VideoJAM is composed of two complementary units. During training, we extend the objective to predict both the generated pixels and their corresponding motion from a single learned representation. During inference, we introduce Inner-Guidance, a mechanism that steers the generation toward coherent motion by leveraging the model's own evolving motion prediction as a dynamic guidance signal. Notably, our framework can be applied to any video model with minimal adaptations, requiring no modifications to the training data or scaling of the model. VideoJAM achieves state-of-the-art performance in motion coherence, surpassing highly competitive proprietary models while also enhancing the perceived visual quality of the generations. These findings emphasize that appearance and motion can be complementary and, when effectively integrated, enhance both the visual quality and the coherence of video generation. Project website: https://hila-chefer.github.io/videojam-paper.github.io/