Enabling High-Level Worker-Centric Semantic Understanding of Onsite Images Using Visual Language Models with Attention Mechanism and Beam Search Strategy

Visual information is becoming increasingly essential in construction management. However, a significant portion of this information remains underutilized by construction managers due to the limitations of existing image processing algorithms. These algorithms primarily rely on low-level visual feat...

Disgrifiad llawn

Manylion Llyfryddiaeth

Prif Awduron:	Hui Deng, Kejie Fu, Binglin Yu, Huimin Li, Rui Duan, Yichuan Deng, Jia-rui Lin
Fformat:	Artigo
Iaith:	Inglês
Cyhoeddwyd:	MDPI AG 2025-03-01
Cyfres:	Buildings
Pynciau:	visual language model construction scene image scene understanding image captioning attention mechanism
Mynediad Ar-lein:	https://www.mdpi.com/2075-5309/15/6/959
Tagiau:	Ychwanegu Tag Dim Tagiau, Byddwch y cyntaf i dagio'r cofnod hwn!