A Survey on Text-Guided 3-D Visual Grounding: Elements, Recent Advances, and Future Directions

Area of Science:

• Computer Vision
• Artificial Intelligence
• Robotics

Background:

• Text-guided 3-D visual grounding (T-3DVG) aims to identify objects in 3-D scenes based on natural language descriptions.
• This task is crucial for human-robot interaction and autonomous systems but faces challenges in data collection and 3-D point cloud processing.

Purpose of the Study:

• To provide a comprehensive survey of the current state-of-the-art in T-3DVG.
• To systematically review fundamental concepts, recent advancements, and future research avenues in the field.
• To offer a tutorial-style overview of the T-3DVG pipeline and its components.

Main Methods:

• Categorization and analysis of existing T-3DVG approaches.
• Review of benchmark datasets and evaluation metrics.
• Discussion of current limitations and future research directions.

Main Results:

• The survey presents a structured overview of T-3DVG, categorizing existing methods and analyzing their performance.
• It highlights the challenges and potential of T-3DVG compared to its 2-D counterpart.
• Key datasets and evaluation metrics are summarized for performance assessment.

Conclusions:

• This work represents the first systematic survey dedicated to T-3DVG.
• It identifies limitations in current T-3DVG research and proposes promising future research directions.
• The survey aims to guide researchers and practitioners in this rapidly evolving field.