English

SPATIALALIGN: Aligning Dynamic Spatial Relationships in Video Generation

Computer Vision and Pattern Recognition 2026-03-02 v2

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. Project page: https://fengming001ntu.github.io/SpatialAlign/

Keywords

Cite

@article{arxiv.2602.22745,
  title  = {SPATIALALIGN: Aligning Dynamic Spatial Relationships in Video Generation},
  author = {Fengming Liu and Tat-Jen Cham and Chuanxia Zheng},
  journal= {arXiv preprint arXiv:2602.22745},
  year   = {2026}
}

Comments

Project page: https://fengming001ntu.github.io/SpatialAlign/

R2 v1 2026-07-01T10:53:30.341Z