3d scene understanding
The capability of an AI system to perceive, interpret, and understand three-dimensional spaces by recognizing objects, their shapes, spatial relationships, and context, critical for applications in robotics, virtual reality, and autonomous navigation.
- EA3D: Online Open-World 3D Object Extraction from Streaming Videos
- Learning from Videos for 3D World: Enhancing MLLMs with 3D Vision Geometry Priors
- MLLM-For3D: Adapting Multimodal Large Language Model for 3D Reasoning Segmentation
- On Geometry-Enhanced Parameter-Efficient Fine-Tuning for 3D Scene Segmentation
- OpenLex3D: A Tiered Benchmark for Open-Vocabulary 3D Scene Representations
- SceneSplat++: A Large Dataset and Comprehensive Benchmark for Language Gaussian Splatting