EmbSpatial-Bench: Benchmarking Spatial Understanding for Embodied Tasks with Large Vision-Language Models

arXiv:2406.05756v1 Announce Type: cross
Abstract: The recent rapid development of Large Vision-Language Models (LVLMs) has indicated their potential for embodied tasks.However, the critical skill of spatial understanding in embodied environments has not been thoroughly evaluated, leaving the gap between current LVLMs and qualified embodied intelligence unknown. Therefore, we construct EmbSpatial-Bench, a benchmark for evaluating embodied spatial understanding of LVLMs.The benchmark is automatically derived from embodied scenes and covers 6 spatial relationships from an egocentric perspective.Experiments expose the insufficient capacity of …

abstract arxiv benchmarking construct cs.ai cs.cl cs.cv cs.mm current development embodied embodied intelligence environments gap however intelligence language language models potential skill spatial tasks type understanding vision vision-language vision-language models