Research

Paper

AI LLM February 26, 2026

SPATIALALIGN: Aligning Dynamic Spatial Relationships in Video Generation

Authors

Fengming Liu, Tat-Jen Cham, Chuanxia Zheng

Abstract

Most text-to-video (T2V) generators prioritize aesthetic quality, but often ignoring the spatial constraints in the generated videos. In this work, we present SPATIALALIGN, a self-improvement framework that enhances T2V models capabilities to depict Dynamic Spatial Relationships (DSR) specified in text prompts. We present a zeroth-order regularized Direct Preference Optimization (DPO) to fine-tune T2V models towards better alignment with DSR. Specifically, we design DSR-SCORE, a geometry-based metric that quantitatively measures the alignment between generated videos and the specified DSRs in prompts, which is a step forward from prior works that rely on VLM for evaluation. We also conduct a dataset of text-video pairs with diverse DSRs to facilitate the study. Extensive experiments demonstrate that our fine-tuned model significantly out performs the baseline in spatial relationships. The code will be released in Link.

Metadata

arXiv ID: 2602.22745
Provider: ARXIV
Primary Category: cs.CV
Published: 2026-02-26
Fetched: 2026-02-27 04:35

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2602.22745v1</id>\n    <title>SPATIALALIGN: Aligning Dynamic Spatial Relationships in Video Generation</title>\n    <updated>2026-02-26T08:34:09Z</updated>\n    <link href='https://arxiv.org/abs/2602.22745v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2602.22745v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Most text-to-video (T2V) generators prioritize aesthetic quality, but often ignoring the spatial constraints in the generated videos. In this work, we present SPATIALALIGN, a self-improvement framework that enhances T2V models capabilities to depict Dynamic Spatial Relationships (DSR) specified in text prompts. We present a zeroth-order regularized Direct Preference Optimization (DPO) to fine-tune T2V models towards better alignment with DSR. Specifically, we design DSR-SCORE, a geometry-based metric that quantitatively measures the alignment between generated videos and the specified DSRs in prompts, which is a step forward from prior works that rely on VLM for evaluation. We also conduct a dataset of text-video pairs with diverse DSRs to facilitate the study. Extensive experiments demonstrate that our fine-tuned model significantly out performs the baseline in spatial relationships. The code will be released in Link.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CV'/>\n    <published>2026-02-26T08:34:09Z</published>\n    <arxiv:primary_category term='cs.CV'/>\n    <author>\n      <name>Fengming Liu</name>\n    </author>\n    <author>\n      <name>Tat-Jen Cham</name>\n    </author>\n    <author>\n      <name>Chuanxia Zheng</name>\n    </author>\n  </entry>"
}