Fire Detection Models Based on Attention Mechanisms and Multiscale Features

Shunxiang Zhang1, Meng Chen1, Kuan-Ching Li2, Hua Wen1 and Liang Sun1

  1. School of Computer Science and Engineering, Anhui University of Science Technology,
    232001 Huainan, China
    sxzhang@aust.edu.cn, 2260662967@qq.com, 1762636707@qq.com, 2582132681@qq.com
  2. Department of Computer Science and Information Engineering (CSIE), 13 Providence
    University, 43301 Taizhong, Taiwan
    kuancli@pu.edu.tw

Abstract

Fire detection is critical in applications such as fire management and building safety, but dispersion and blurring of flame and smoke boundaries can present challenges. Multiple upsampling and downsampling operations can blur the localisation signals, thus reducing accuracy and efficiency. To address this problem, we propose the AMMF(Attention Mechanisms and Multiscale Features) detection model, which integrates an attention mechanism and multi-scale feature fusion to improve accuracy and real-time performance. The model incorporates a dynamic sparse attention mechanism in the backbone network to enhance feature capture and restructures the neck network using CepBlock and MPFusion modules for better feature fusion. MDPIoU loss and Slideloss are then utilised to reduce the bounding box regression error and address the sample imbalance problem respectively. In addition, parameters are shared by merging 3×3 convolutional branches, which optimises the detection head and improves computational efficiency. The experimental results show that AMMF-Detection can significantly improve the detection speed and accuracy on the public dataset.

Key words

Fire detection,YOLO,Feature fusion,ynamic sparse attention,Multi-scale features

Digital Object Identifier (DOI)

https://doi.org/10.2298/CSIS241225059Z

How to cite

Zhang, S., Chen, M., Li, K., Wen, H., Sun, L.: Fire Detection Models Based on Attention Mechanisms and Multiscale Features. Computer Science and Information Systems, https://doi.org/10.2298/CSIS241225059Z