Comparison of Spectrograms for Classification of Vehicles from Traffic Audio
Traditional vision-based traffic monitoring struggles with cost and low-light conditions. This research explores audio-based vehicle classification using deep learning to address these limitations. This exploration investigates the impact of different spectrogram types namely Short-time Fourier Tran...
Published in: | IEEE Symposium on Wireless Technology and Applications, ISWTA |
---|---|
Main Author: | |
Format: | Conference paper |
Language: | English |
Published: |
IEEE Computer Society
2024
|
Online Access: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-85203799017&doi=10.1109%2fISWTA62130.2024.10651848&partnerID=40&md5=29d22ec0b8c2c502aa2ee0af5e1a00eb |
id |
2-s2.0-85203799017 |
---|---|
spelling |
2-s2.0-85203799017 Shariff K.K.M.; Haron M.A.; Ali M.S.A.M.; Yassin I.M.; Azami M.H.; Kechik M.A. Comparison of Spectrograms for Classification of Vehicles from Traffic Audio 2024 IEEE Symposium on Wireless Technology and Applications, ISWTA 10.1109/ISWTA62130.2024.10651848 https://www.scopus.com/inward/record.uri?eid=2-s2.0-85203799017&doi=10.1109%2fISWTA62130.2024.10651848&partnerID=40&md5=29d22ec0b8c2c502aa2ee0af5e1a00eb Traditional vision-based traffic monitoring struggles with cost and low-light conditions. This research explores audio-based vehicle classification using deep learning to address these limitations. This exploration investigates the impact of different spectrogram types namely Short-time Fourier Transform (STFT), Mel, Gammatonegram, Constant-Q, Wavelet and their combinations on classification accuracy using a Convolutional Neural Network based on AlexNet on the IDMT-traffic dataset. Results showed that Individual spectrogram performance varied by vehicle class, with STFT excelling for cars and no-traffic, and GAM for motorcycles and trucks. Furthermore, combining spectrograms yielded some slight improvements (up to 3%), but also occasional drops, likely due to feature loss during fusion. Our study highlights the importance of choosing the right spectrogram type based on the target vehicle class. While simple combinations showed limited improvements, exploring more sophisticated fusion techniques holds promise for further enhancing audio-based vehicle classification accuracy. © 2024 IEEE. IEEE Computer Society 23247843 English Conference paper |
author |
Shariff K.K.M.; Haron M.A.; Ali M.S.A.M.; Yassin I.M.; Azami M.H.; Kechik M.A. |
spellingShingle |
Shariff K.K.M.; Haron M.A.; Ali M.S.A.M.; Yassin I.M.; Azami M.H.; Kechik M.A. Comparison of Spectrograms for Classification of Vehicles from Traffic Audio |
author_facet |
Shariff K.K.M.; Haron M.A.; Ali M.S.A.M.; Yassin I.M.; Azami M.H.; Kechik M.A. |
author_sort |
Shariff K.K.M.; Haron M.A.; Ali M.S.A.M.; Yassin I.M.; Azami M.H.; Kechik M.A. |
title |
Comparison of Spectrograms for Classification of Vehicles from Traffic Audio |
title_short |
Comparison of Spectrograms for Classification of Vehicles from Traffic Audio |
title_full |
Comparison of Spectrograms for Classification of Vehicles from Traffic Audio |
title_fullStr |
Comparison of Spectrograms for Classification of Vehicles from Traffic Audio |
title_full_unstemmed |
Comparison of Spectrograms for Classification of Vehicles from Traffic Audio |
title_sort |
Comparison of Spectrograms for Classification of Vehicles from Traffic Audio |
publishDate |
2024 |
container_title |
IEEE Symposium on Wireless Technology and Applications, ISWTA |
container_volume |
|
container_issue |
|
doi_str_mv |
10.1109/ISWTA62130.2024.10651848 |
url |
https://www.scopus.com/inward/record.uri?eid=2-s2.0-85203799017&doi=10.1109%2fISWTA62130.2024.10651848&partnerID=40&md5=29d22ec0b8c2c502aa2ee0af5e1a00eb |
description |
Traditional vision-based traffic monitoring struggles with cost and low-light conditions. This research explores audio-based vehicle classification using deep learning to address these limitations. This exploration investigates the impact of different spectrogram types namely Short-time Fourier Transform (STFT), Mel, Gammatonegram, Constant-Q, Wavelet and their combinations on classification accuracy using a Convolutional Neural Network based on AlexNet on the IDMT-traffic dataset. Results showed that Individual spectrogram performance varied by vehicle class, with STFT excelling for cars and no-traffic, and GAM for motorcycles and trucks. Furthermore, combining spectrograms yielded some slight improvements (up to 3%), but also occasional drops, likely due to feature loss during fusion. Our study highlights the importance of choosing the right spectrogram type based on the target vehicle class. While simple combinations showed limited improvements, exploring more sophisticated fusion techniques holds promise for further enhancing audio-based vehicle classification accuracy. © 2024 IEEE. |
publisher |
IEEE Computer Society |
issn |
23247843 |
language |
English |
format |
Conference paper |
accesstype |
|
record_format |
scopus |
collection |
Scopus |
_version_ |
1812871795855851520 |