View in EDS

Surface Fire Test Using Diffusion-Augmented Multimodal Deep Belief Learning.

Saved in:
Bibliographic Details
Title: Surface Fire Test Using Diffusion-Augmented Multimodal Deep Belief Learning.
Authors: Aurnob, Al Rafi1 (AUTHOR) inc.aurnob@gmail.com, Tanim, Sharia Arfin1 (AUTHOR) shariaarfin096@gmail.com, Biswas, Tapu2 (AUTHOR) tapubiswas731@gmail.com, Mridha, M. F.1 (AUTHOR) firoz.mridha@aiub.edu, Safran, Mejdl3 (AUTHOR) mejdl@ksu.edu.sa, Alfarhood, Sultan3 (AUTHOR) sultanf@ksu.edu.sa, Che, Dunren4 (AUTHOR) dunren.che@tamuk.edu
Source: Fire Technology. May2026, Vol. 62 Issue 3, p1-34. 34p.
Abstract: One of the issues with a wildfire detection system is its limited and poor-quality datasets of images, which hinder the performance of accurate and timely detection. To address this, as a solution, the proposed study aims to utilize a Generative Artificial Intelligence (GAI) based methodology, in which high-fidelity wildfire-related imagery would be generated using Stable Diffusion, thus improving the quality and level of training data diversity. A multimodal deep belief network (DBN) is used for classification using deep learning capabilities to improve the accuracy of distinguishing between fire and non-fire scenarios. Explainable Artificial Intelligence (XAI) techniques are incorporated into the classification framework as a further attempt to increase the transparency of the model. The experimental results demonstrate the effectiveness of the proposed approach, achieving the test accuracy of 99.67% in the synthetic data set and 99.55% and 99.63% in two real-world data sets. The results show that the integration of generative artificial intelligence, deep learning, and explainability contributes to better wildfire detection performance. Furthermore, this method delivers more efficient and transparent AI tools, which could be trained in various settings, and can support the creation of high-quality contextual-adequate imagery. [ABSTRACT FROM AUTHOR]
Copyright of Fire Technology is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
Description
Abstract:One of the issues with a wildfire detection system is its limited and poor-quality datasets of images, which hinder the performance of accurate and timely detection. To address this, as a solution, the proposed study aims to utilize a Generative Artificial Intelligence (GAI) based methodology, in which high-fidelity wildfire-related imagery would be generated using Stable Diffusion, thus improving the quality and level of training data diversity. A multimodal deep belief network (DBN) is used for classification using deep learning capabilities to improve the accuracy of distinguishing between fire and non-fire scenarios. Explainable Artificial Intelligence (XAI) techniques are incorporated into the classification framework as a further attempt to increase the transparency of the model. The experimental results demonstrate the effectiveness of the proposed approach, achieving the test accuracy of 99.67% in the synthetic data set and 99.55% and 99.63% in two real-world data sets. The results show that the integration of generative artificial intelligence, deep learning, and explainability contributes to better wildfire detection performance. Furthermore, this method delivers more efficient and transparent AI tools, which could be trained in various settings, and can support the creation of high-quality contextual-adequate imagery. [ABSTRACT FROM AUTHOR]
ISSN:00152684
DOI:10.1007/s10694-026-01881-8