Automated accurate fire detection system using ensemble pretrained residual network

作者：

Highlights：

• Four pretrained ResNets are utilized to create learning models.

• Two ensemble learning models are proposed using two approximations.

• Our models are tested on a fire image dataset.

• Our models attained 98.91% and 99.15% accuracies on the used dataset.

摘要

•Four pretrained ResNets are utilized to create learning models.•Two ensemble learning models are proposed using two approximations.•Our models are tested on a fire image dataset.•Our models attained 98.91% and 99.15% accuracies on the used dataset.

论文关键词：Fire detection,Ensemble ResNet,Deep feature extraction,Transfer learning,Iterative hard majority voting,NCA

论文评审过程：Received 11 November 2021, Revised 18 January 2022, Accepted 25 April 2022, Available online 2 May 2022, Version of Record 10 May 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.117407