SpatialBot: Precise Spatial Understanding with Vision Language Models

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Cai, Wenxiao, Ponomarenko, Iaroslav, Yuan, Jianhao, Li, Xiaoqi, Yang, Wankou, Dong, Hao, Zhao, Bo
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866909541540560896
author Cai, Wenxiao
Ponomarenko, Iaroslav
Yuan, Jianhao
Li, Xiaoqi
Yang, Wankou
Dong, Hao
Zhao, Bo
author_facet Cai, Wenxiao
Ponomarenko, Iaroslav
Yuan, Jianhao
Li, Xiaoqi
Yang, Wankou
Dong, Hao
Zhao, Bo
contents Vision Language Models (VLMs) have achieved impressive performance in 2D image understanding, however they are still struggling with spatial understanding which is the foundation of Embodied AI. In this paper, we propose SpatialBot for better spatial understanding by feeding both RGB and depth images. Additionally, we have constructed the SpatialQA dataset, which involves multi-level depth-related questions to train VLMs for depth understanding. Finally, we present SpatialBench to comprehensively evaluate VLMs' capabilities in spatial understanding at different levels. Extensive experiments on our spatial-understanding benchmark, general VLM benchmarks and Embodied AI tasks, demonstrate the remarkable improvements of SpatialBot trained on SpatialQA. The model, code and data are available at https://github.com/BAAI-DCAI/SpatialBot.
format Preprint
id arxiv_https___arxiv_org_abs_2406_13642
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SpatialBot: Precise Spatial Understanding with Vision Language Models
Cai, Wenxiao
Ponomarenko, Iaroslav
Yuan, Jianhao
Li, Xiaoqi
Yang, Wankou
Dong, Hao
Zhao, Bo
Computer Vision and Pattern Recognition
Vision Language Models (VLMs) have achieved impressive performance in 2D image understanding, however they are still struggling with spatial understanding which is the foundation of Embodied AI. In this paper, we propose SpatialBot for better spatial understanding by feeding both RGB and depth images. Additionally, we have constructed the SpatialQA dataset, which involves multi-level depth-related questions to train VLMs for depth understanding. Finally, we present SpatialBench to comprehensively evaluate VLMs' capabilities in spatial understanding at different levels. Extensive experiments on our spatial-understanding benchmark, general VLM benchmarks and Embodied AI tasks, demonstrate the remarkable improvements of SpatialBot trained on SpatialQA. The model, code and data are available at https://github.com/BAAI-DCAI/SpatialBot.
title SpatialBot: Precise Spatial Understanding with Vision Language Models
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2406.13642