Summary

This repository serves as a curated, up-to-date collection of papers, methods, and resources for Text-guided 3D Visual Grounding (T-3DVG). It addresses the challenge of locating objects in complex 3D scenes based on natural language queries, benefiting researchers and engineers by providing a centralized overview of the rapidly evolving field.

How It Works

The project functions as a comprehensive, categorized bibliography of T-3DVG research. It systematically organizes methods into distinct approaches, including fully-supervised (one-stage and two-stage), weakly-supervised, semi-supervised, LLM-based, and outdoor scene-focused techniques. Each entry typically links to the corresponding research paper, dataset, and often a GitHub repository for implementation details.

Highlighted Details

• Features a detailed categorization of T-3DVG methods, aiding systematic literature review.
• Includes emerging trends such as LLM integration and applications in autonomous driving/outdoor scenes.
• Provides direct links to academic papers, datasets, and code repositories for practical exploration.
• Highlights a recent survey paper offering a deep dive into T-3DVG elements, advances, and future directions.

Maintenance & Community

The list is maintained through community contributions, with contact information provided for submitting missed works (daizongliu@whu.edu.cn). The primary contribution is the curated collection itself, referencing significant academic publications.

Licensing & Compatibility

No specific license is provided for this curated list. Individual papers and code repositories linked within the list will have their own respective licenses, which users must consult for usage restrictions.

Limitations & Caveats

This repository is a reference guide, not a deployable software tool. Its utility is dependent on the quality and accessibility of the linked external research papers and code. The "up-to-date" status is relative to the pace of research publication in this active domain.