QR-koodi

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...

Täydet tiedot

Tallennettuna:
Päätekijät: Hui Deng, Kejie Fu, Binglin Yu, Huimin Li, Rui Duan, Yichuan Deng, Jia-rui Lin
Aineistotyyppi: Artigo
Kieli:Inglês
Julkaistu: MDPI AG 2025-03-01
Sarja:Buildings
Aiheet:
visual language model
construction scene
image scene understanding
image captioning
attention mechanism
Linkit:https://www.mdpi.com/2075-5309/15/6/959
Tagit: Lisää tagi
Ei tageja, Lisää ensimmäinen tagi!

Samankaltaisia teoksia