Skip to Content

Pretrained Neural Network Models for Thomson 2020 study - PyTorch format Open Access

Automatic visual fire detection is used to complement traditional fire detection sensor systems (smoke/heat). In this work, we investigate different Convolutional Neural Network (CNN) architectures and their variants for the nontemporal real-time bounds detection of fire pixel regions in video (or still) imagery. Two reduced complexity compact CNN architectures (NasNet-A-OnFire and ShuffleNetV2-OnFire) are proposed through experimental analysis to optimise the computational efficiency for this task. The results improve upon the current state-of-the-art solution for fire detection, achieving an accuracy of 95% for full-frame binary classification and 97% for superpixel localisation. We notably achieve a classification speed up by a factor of 2.3× for binary classification and 1.3× for superpixel localisation, with runtime of 40 fps and 18 fps respectively, outperforming prior work in the field presenting an efficient, robust and real-time solution for fire region detection. Subsequent implementation on low-powered devices (Nvidia Xavier-NX, achieving 49 fps for full-frame classification via ShuffleNetV2-OnFire) demonstrates our architectures are suitable for various real-world deployment applications. | Cited in: Efficient and Compact Convolutional Neural Network Architectures for Non-temporal Real-time Fire Detection (W. Thomson, N. Bhowmik, T.P. Breckon), In Proc. Int. Conf. on Machine Learning Applications, IEEE, 2020. | This file contains supporting materials in the form of the pre-trained network models used in the study.

Descriptions

Resource type
Other
Contributors
Creator: Thomson, William 1
Contact person: Bhowmik, Neelanjan 1
Editor: Bhowmik, Neelanjan 1
Editor: Breckon, Toby P. 1
1 Durham University, UK
Funder
Durham University
Research methods
Other description
Keyword
Convolutional Neural Network
fire detection
classification
Subject
Engineering
Computer science
Location
Durham, England, United Kingdom
Language
English
Cited in
Efficient and Compact Convolutional Neural Network Architectures for Non-temporal Real-time Fire Detection (W. Thomson, N. Bhowmik, T.P. Breckon), In Proc. Int. Conf. on Machine Learning Applications, IEEE, 2020
Identifier
ark:/32150/r1tb09j570z
doi:10.15128/r1tb09j570z
Rights
MIT Licence (MIT)

Creative Commons Attribution 4.0 International (CC BY)

Publisher
Durham University
Date Created

File Details

Depositor
N. Bhowmik
Date Uploaded
Date Modified
7 March 2021, 19:03:43
Audit Status
Audits have not yet been run on this file.
Characterization
File format: zip (ZIP Format)
Mime type: application/zip
File size: 155995759
Last modified: 2021:03:07 19:17:00+00:00
Filename: thomson-2020-fire-detection-pretrained-models-pytorch--version2.zip
Original checksum: 5c7be76d7f8149be4057849c998ab214
Activity of users you follow
User Activity Date
User N. Bhowmik has updated Pretrained Neural Network Models for Thomson 2020 study - PyTorch format about 3 years ago
User N. Bhowmik has updated thomson-2020-fire-detection-pretrained-models-pytorch--version1.zip about 3 years ago
User N. Bhowmik has updated thomson-2020-fire-detection-pretrained-models-pytorch--version1.zip about 3 years ago
User N. Bhowmik has deposited thomson-2020-fire-detection-pretrained-models-pytorch--version1.zip about 3 years ago