Experimental Exploration of Compact Convolutional Neural Network Architectures for Non-temporal Real-time Fire Detection

TL;DR

This paper investigates reduced complexity CNN architectures for real-time fire detection in images and videos, achieving high accuracy and low false positive rates with efficient models based on InceptionV4.

Contribution

It introduces experimentally optimized, lightweight CNN architectures for fire detection that outperform prior models in accuracy and false positive rate.

Findings

Maximum accuracy of 0.96 for fire detection

False positive rate of 0.06, lower than previous work

Effective use of reduced CNN architectures based on InceptionV4

Abstract

In this work we explore different Convolutional Neural Network (CNN) architectures and their variants for non-temporal binary fire detection and localization in video or still imagery. We consider the performance of experimentally defined, reduced complexity deep CNN architectures for this task and evaluate the effects of different optimization and normalization techniques applied to different CNN architectures (spanning the Inception, ResNet and EfficientNet architectural concepts). Contrary to contemporary trends in the field, our work illustrates a maximum overall accuracy of 0.96 for full frame binary fire detection and 0.94 for superpixel localization using an experimentally defined reduced CNN architecture based on the concept of InceptionV4. We notably achieve a lower false positive rate of 0.06 compared to prior work in the field presenting an efficient, robust and real-time solution for fire region detection.