Through-The-Mask: Mask-based Motion Trajectories for Image-to-Video Generation

TL;DR

This paper introduces a mask-based motion trajectory representation for image-to-video generation, significantly improving motion accuracy and temporal coherence in multi-object scenarios by decomposing the task into intermediate representation generation and video synthesis.

Contribution

The paper proposes a novel two-stage framework with a mask-based motion trajectory as an intermediate representation, enhancing motion realism and consistency in image-to-video generation.

Findings

Achieves state-of-the-art results in temporal coherence and motion realism.

Introduces a new benchmark for multi-object image-to-video generation.

Demonstrates superior performance on challenging benchmarks.

Abstract

We consider the task of Image-to-Video (I2V) generation, which involves transforming static images into realistic video sequences based on a textual description. While recent advancements produce photorealistic outputs, they frequently struggle to create videos with accurate and consistent object motion, especially in multi-object scenarios. To address these limitations, we propose a two-stage compositional framework that decomposes I2V generation into: (i) An explicit intermediate representation generation stage, followed by (ii) A video generation stage that is conditioned on this representation. Our key innovation is the introduction of a mask-based motion trajectory as an intermediate representation, that captures both semantic object information and motion, enabling an expressive but compact representation of motion and semantics. To incorporate the learned representation in the second stage, we utilize object-level attention objectives. Specifically, we consider a spatial, per-object, masked-cross attention objective, integrating object-specific prompts into corresponding latent space regions and a masked spatio-temporal self-attention objective, ensuring frame-to-frame consistency for each object. We evaluate our method on challenging benchmarks with multi-object and high-motion scenarios and empirically demonstrate that the proposed method achieves state-of-the-art results in temporal coherence, motion realism, and text-prompt faithfulness. Additionally, we introduce \benchmark, a new challenging benchmark for single-object and multi-object I2V generation, and demonstrate our method's superiority on this benchmark. Project page is available at https://guyyariv.github.io/TTM/.