Training-Free Learning

CitySTAR: Agent-Driven Structured and Topology-Aware Reasoning for Open-Vocabulary Urban 3D Grounding featured image
Open-Vocabulary 3D Grounding

CitySTAR: Agent-Driven Structured and Topology-Aware Reasoning for Open-Vocabulary Urban 3D Grounding

NeurIPS 2026 · First author. CitySTAR grounds natural-language queries in city-scale 3D scenes through structured constraints, scene graphs and topology-aware reasoning.

shuai-zhang
•
Read more
Open-Vocabulary 3D Segmentation

GeoOpen3D: Geometry-guided training-free open-vocabulary 3D segmentation via visual foundation models

Geometry-guided RGB-D projection and visual foundation models enable training-free open-vocabulary 3D point-cloud segmentation.

shuai-zhang
•
Read more