Implementing an Efficient Speed Bump Detection System Using Adaptive Threshold Gaussian over Support Vector Machine for Improved Detection

R. Priyanka, W. Priya

2023

Abstract

This research endeavours to identify speed bumps from provided images using Adaptive Thresholding for enhanced detection. A total of 120 samples were divided equally into two groups. The first group, comprising 60 samples, underwent testing using the Support Vector Machine, while the second group was tested with the Adaptive Threshold-Gaussian. Each group underwent 10 iterations. The dataset, comprising 6000 images sourced from Kaggle.com, allocated 4800 images for training and the remaining for testing. With a G power roughly at 80%, the Gaussian Adaptive Threshold yielded an accuracy of 85.60%, surpassing the Support Vector Machine's 81.40%. A significance value of 0.002 (p<0.05) indicates that the results between the two groups are statistically significant. The Gaussian Adaptive Threshold, therefore, stands out for its superior accuracy.

Paper Citation

in Harvard Style

Priyanka R. and Priya W. (2023). Implementing an Efficient Speed Bump Detection System Using Adaptive Threshold Gaussian over Support Vector Machine for Improved Detection. In Proceedings of the 1st International Conference on Artificial Intelligence for Internet of Things: Accelerating Innovation in Industry and Consumer Electronics - Volume 1: AI4IoT; ISBN 978-989-758-661-3, SciTePress, pages 527-532. DOI: 10.5220/0012603100003739

in Bibtex Style

@conference{ai4iot23,
author={R. Priyanka and W. Priya},
title={Implementing an Efficient Speed Bump Detection System Using Adaptive Threshold Gaussian over Support Vector Machine for Improved Detection},
booktitle={Proceedings of the 1st International Conference on Artificial Intelligence for Internet of Things: Accelerating Innovation in Industry and Consumer Electronics - Volume 1: AI4IoT},
year={2023},
pages={527-532},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012603100003739},
isbn={978-989-758-661-3},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Artificial Intelligence for Internet of Things: Accelerating Innovation in Industry and Consumer Electronics - Volume 1: AI4IoT
TI - Implementing an Efficient Speed Bump Detection System Using Adaptive Threshold Gaussian over Support Vector Machine for Improved Detection
SN - 978-989-758-661-3
AU - Priyanka R.
AU - Priya W.
PY - 2023
SP - 527
EP - 532
DO - 10.5220/0012603100003739
PB - SciTePress