---
title: "Syntactic Structure-Guided Visual Grounding with Subject-Centric Feature Enhancement and Verification"
authors:
  - "Jiepeng Cai"
  - "Zhen Xu"
  - "Tiesong Zhao"
  - "Hau-San Wong"
  - "Si Wu"
doi: "10.24963/ijcai.2026/104"
url: "https://ijcai.org/proceedings/2026/104"
pdf: "https://ijcai.org/proceedings/2026/0104.pdf"
venue: "Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence"
year: 2026
pages: "926-934"
track: "Main Track"
keywords:
  - "Computer Vision: Segmentation, grouping and shape analysis"
  - "Computer Vision: Multimodal learning"
  - "Computer Vision: Recognition (object detection, categorization)"
---

# Syntactic Structure-Guided Visual Grounding with Subject-Centric Feature Enhancement and Verification

Jiepeng Cai, Zhen Xu, Tiesong Zhao, Hau-San Wong, Si Wu

*Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence*, pages 926-934. DOI: [10.24963/ijcai.2026/104](https://doi.org/10.24963/ijcai.2026/104)

## Abstract

Visual grounding aims to localize target objects based on natural language descriptions, and the core challenge lies in the cross-modal gap, which is partly caused by the significant differences in semantic structure between language and vision. Existing methods typically rely on holistic sentence-level semantic representations to modulate visual features, while overlooking the inherent structure of textual prompts. In this work, we propose a Syntactic Structure-guided Visual Grounding framework, referred to as SSVG. Specifically, to inject syntactic priors into unstructured visual representations, we design a semantic structure-based feature refinement module to adaptively modulate subject-centric and contextual visual features. To perform cross-modal alignment, we further incorporate a visual semantic consistency verification module, which leverages a subject-aware contrastive learning strategy to constrain and verify the semantic correspondence between the visual prediction and subject-level textual representation, thereby enhancing the model's robustness against semantically similar distractors. Extensive experiments demonstrate that our approach consistently outperforms state-of-the-art methods, and detailed analyses verify the effectiveness of each component.