Research Article
Image Classification for Smoke and Flame Recognition Using CNN and Transfer Learning on Edge Device
@INPROCEEDINGS{10.1007/978-3-031-31275-5_16, author={Endah Kristiani and Yi-Chun Chen and Chao-Tung Yang and Chia-Hsin Li}, title={Image Classification for Smoke and Flame Recognition Using CNN and Transfer Learning on Edge Device}, proceedings={Smart Grid and Internet of Things. 6th EAI International Conference, SGIoT 2022, TaiChung, Taiwan, November 19-20, 2022, Proceedings}, proceedings_a={SGIOT}, year={2023}, month={5}, keywords={image classification CNN Imagnet edge computing transfer learning}, doi={10.1007/978-3-031-31275-5_16} }- Endah Kristiani
Yi-Chun Chen
Chao-Tung Yang
Chia-Hsin Li
Year: 2023
Image Classification for Smoke and Flame Recognition Using CNN and Transfer Learning on Edge Device
SGIOT
Springer
DOI: 10.1007/978-3-031-31275-5_16
Abstract
This paper implemented image classification for smoke and flame detection. CNN model was trained in three topologies of InceptionV3, MobileNet, and VGG16. These three models were then tested on Raspberry Pi 4 with Intel Neural Compute Stick 2 (NCS 2). The experimental results demonstrated that MobileNetV2 is a superior model to the other two models in terms of training and inference, even if the accuracy rate of the three was as high as 94% when utilizing the test set for evaluation.
Copyright © 2022–2025 ICST
