Drone Detection with Improved Precision in Traditional Machine Learning and Less Complexity in Single Shot Detectors

This work presents a broad study of drone detection based on a variety of machine-learning methods including traditional and deep-learning techniques. The data sets used are images obtained from sequences of video frames in both RGB and IR formats, filtered and unfiltered. First, traditional machine...

Full description

Saved in:
Bibliographic Details
Main Author: Kassab, Mohamad (author)
Other Authors: Abu Zitar, Raed (author), Barbaresco, Frederic (author), Seghrouchni, Amal El Fallah (author)
Published: 2024
Subjects:
Online Access:https://depot.sorbonne.ae/handle/20.500.12458/1480
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1857415064866258944
author Kassab, Mohamad
author2 Abu Zitar, Raed
Barbaresco, Frederic
Seghrouchni, Amal El Fallah
author2_role author
author
author
author_facet Kassab, Mohamad
Abu Zitar, Raed
Barbaresco, Frederic
Seghrouchni, Amal El Fallah
author_role author
dc.creator.none.fl_str_mv Kassab, Mohamad
Abu Zitar, Raed
Barbaresco, Frederic
Seghrouchni, Amal El Fallah
dc.date.none.fl_str_mv 2024-02-27T06:14:27Z
2024-02-27T06:14:27Z
2024
dc.identifier.none.fl_str_mv 0018-9251
1557-9603
2371-9877
https://depot.sorbonne.ae/handle/20.500.12458/1480
10.1109/TAES.2024.3368991
dc.language.none.fl_str_mv en
dc.relation.none.fl_str_mv IEEE Transactions on Aerospace and Electronic Systems
dc.subject.none.fl_str_mv Drone Detection
Deep Learning
AdderNet
Computational Complexity
Traditional Machine Learning
dc.title.none.fl_str_mv Drone Detection with Improved Precision in Traditional Machine Learning and Less Complexity in Single Shot Detectors
dc.type.none.fl_str_mv Controlled Vocabulary for Resource Type Genres::text::periodical::journal::contribution to journal::journal article
description This work presents a broad study of drone detection based on a variety of machine-learning methods including traditional and deep-learning techniques. The data sets used are images obtained from sequences of video frames in both RGB and IR formats, filtered and unfiltered. First, traditional machine learning techniques such as SVM and RF were investigated to discover their drawbacks and study their feasibility in drone detection. It was evident that those techniques are not suitable for complex data sets (sets with several non-drone objects and clutter in the background). It was observed that the sliding window size results in a bias toward the selection of the bounding box when using the traditional NMS method. Therefore, to address this issue, a modified NMS is proposed and tested on SVM and RF. SVM and RF with modified NMS managed to achieve a relative improvement of up to 25% based on the evaluation metric. The Deep Learning techniques, on the other hand, showed better detection performance but less improvement when using the proposed NMS method. Since their biggest drawback is complexity, a modified deep learning paradigm was proposed to mitigate the usual complexity associated with deep learning methods. The proposed paradigm uses (Single Shot Detector) SSD and AdderNet filters in an attempt to avoid excessive multiplications in the convolutional layers. To demonstrate our method, the most common deep-learning techniques were comparatively tested to create a baseline for evaluating the proposed SSD/AdderNet. The training and testing of the deep learning models were repeated six times to investigate the consistency of learning in terms of parameters and performance. The proposed model was able to achieve better results with respect to the IR data set compared to its counterpart while reducing the number of multiplications at the convolutional layers by 43.42%. Moreover, and as a result of lower complexity, the proposed SSD/AdderNet showed fewer training and inference times compared to its counterpart.
id sorbonner_57bdf74e0a56ecaa8fd7b79266567bed
identifier_str_mv 0018-9251
1557-9603
2371-9877
10.1109/TAES.2024.3368991
language_invalid_str_mv en
network_acronym_str sorbonner
network_name_str Sorbonne University Abu Dhabi repository
oai_identifier_str oai:depot.sorbonne.ae:20.500.12458/1480
publishDate 2024
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
spelling Drone Detection with Improved Precision in Traditional Machine Learning and Less Complexity in Single Shot DetectorsKassab, MohamadAbu Zitar, RaedBarbaresco, FredericSeghrouchni, Amal El FallahDrone DetectionDeep LearningAdderNetComputational ComplexityTraditional Machine LearningThis work presents a broad study of drone detection based on a variety of machine-learning methods including traditional and deep-learning techniques. The data sets used are images obtained from sequences of video frames in both RGB and IR formats, filtered and unfiltered. First, traditional machine learning techniques such as SVM and RF were investigated to discover their drawbacks and study their feasibility in drone detection. It was evident that those techniques are not suitable for complex data sets (sets with several non-drone objects and clutter in the background). It was observed that the sliding window size results in a bias toward the selection of the bounding box when using the traditional NMS method. Therefore, to address this issue, a modified NMS is proposed and tested on SVM and RF. SVM and RF with modified NMS managed to achieve a relative improvement of up to 25% based on the evaluation metric. The Deep Learning techniques, on the other hand, showed better detection performance but less improvement when using the proposed NMS method. Since their biggest drawback is complexity, a modified deep learning paradigm was proposed to mitigate the usual complexity associated with deep learning methods. The proposed paradigm uses (Single Shot Detector) SSD and AdderNet filters in an attempt to avoid excessive multiplications in the convolutional layers. To demonstrate our method, the most common deep-learning techniques were comparatively tested to create a baseline for evaluating the proposed SSD/AdderNet. The training and testing of the deep learning models were repeated six times to investigate the consistency of learning in terms of parameters and performance. The proposed model was able to achieve better results with respect to the IR data set compared to its counterpart while reducing the number of multiplications at the convolutional layers by 43.42%. Moreover, and as a result of lower complexity, the proposed SSD/AdderNet showed fewer training and inference times compared to its counterpart.2024-02-27T06:14:27Z2024-02-27T06:14:27Z2024Controlled Vocabulary for Resource Type Genres::text::periodical::journal::contribution to journal::journal article0018-92511557-96032371-9877https://depot.sorbonne.ae/handle/20.500.12458/148010.1109/TAES.2024.3368991enIEEE Transactions on Aerospace and Electronic Systemsoai:depot.sorbonne.ae:20.500.12458/14802024-02-27T06:14:27Z
spellingShingle Drone Detection with Improved Precision in Traditional Machine Learning and Less Complexity in Single Shot Detectors
Kassab, Mohamad
Drone Detection
Deep Learning
AdderNet
Computational Complexity
Traditional Machine Learning
title Drone Detection with Improved Precision in Traditional Machine Learning and Less Complexity in Single Shot Detectors
title_full Drone Detection with Improved Precision in Traditional Machine Learning and Less Complexity in Single Shot Detectors
title_fullStr Drone Detection with Improved Precision in Traditional Machine Learning and Less Complexity in Single Shot Detectors
title_full_unstemmed Drone Detection with Improved Precision in Traditional Machine Learning and Less Complexity in Single Shot Detectors
title_short Drone Detection with Improved Precision in Traditional Machine Learning and Less Complexity in Single Shot Detectors
title_sort Drone Detection with Improved Precision in Traditional Machine Learning and Less Complexity in Single Shot Detectors
topic Drone Detection
Deep Learning
AdderNet
Computational Complexity
Traditional Machine Learning
url https://depot.sorbonne.ae/handle/20.500.12458/1480