SpatialLock: Precise Spatial Control in Text-to-Image Synthesis

TL;DR

SpatialLock introduces a novel framework for text-to-image synthesis that significantly improves object localization accuracy by leveraging perception signals and grounding information, enabling precise spatial control and higher visual quality.

Contribution

The paper presents SpatialLock, a new method combining position-guided learning and attention-based injection to enhance spatial control in text-to-image generation.

Findings

Achieves IOU scores above 0.9 across multiple datasets.

Sets a new state-of-the-art for object positioning accuracy.

Improves visual quality of generated images.

Abstract

Text-to-Image (T2I) synthesis has made significant advancements in recent years, driving applications such as generating datasets automatically. However, precise control over object localization in generated images remains a challenge. Existing methods fail to fully utilize positional information, leading to an inadequate understanding of object spatial layouts. To address this issue, we propose SpatialLock, a novel framework that leverages perception signals and grounding information to jointly control the generation of spatial locations. SpatialLock incorporates two components: Position-Engaged Injection (PoI) and Position-Guided Learning (PoG). PoI directly integrates spatial information through an attention layer, encouraging the model to learn the grounding information effectively. PoG employs perception-based supervision to further refine object localization. Together, these components enable the model to generate objects with precise spatial arrangements and improve the visual quality of the generated images. Experiments show that SpatialLock sets a new state-of-the-art for precise object positioning, achieving IOU scores above 0.9 across multiple datasets.