A Pedestrian Detection Case Study for a Traffic Light Controller

Alexander Wendt; Horst Possegger; Matthias Bittner; Daniel Schnöll; Matthias Wess; Dušan Malić; Horst Bischof; Axel Jantsch

doi:10.1007/978-3-031-39932-9_4

A Pedestrian Detection Case Study for a Traffic Light Controller

Alexander Wendt, Horst Possegger, Matthias Bittner, Daniel Schnöll, Matthias Wess, Dušan Malić, Horst Bischof, Axel Jantsch^*

^*Korrespondierende/r Autor/-in für diese Arbeit

Institut für Maschinelles Sehen und Darstellen (7100)

Publikation: Beitrag in Buch/Bericht/Konferenzband › Beitrag in Buch/Bericht › Begutachtung

Abstract

A pedestrian detection system in a traffic light controller is studied. The system is based on Deep Neural Networks (DNNs). We explore several network architectures and hardware platforms to identify the most suitable solution under the given constraints of latency, cost, and precision. Specifically, we study altogether 13 networks from the MobileNet, Yolo, ResNet, and EfficientDet families and 6 platforms based on Nvidia and Intel platforms, conducting 383 experiments. We find that several network-platform combinations meet the given requirements of maximum 100 ms inference latency and 0.9 mean average precision. The most promising are Yolo v5 networks on Nvidia Jetson TX2 and IntelNUC GPU hardware.

Originalsprache	englisch
Titel	Embedded Machine Learning for Cyber-Physical, IoT, and Edge Computing
Untertitel	Software Optimazations and Hardware/Software Codesign
Redakteure/-innen	Sudeep Pasricha, Muhammad Shafique
Erscheinungsort	Cham
Herausgeber (Verlag)	Springer
Seiten	75-96
ISBN (Print)	978-3-031-39931-2
DOIs	https://doi.org/10.1007/978-3-031-39932-9_4
Publikationsstatus	Veröffentlicht - 2023

Zugriff auf Dokument

10.1007/978-3-031-39932-9_4

Dieses zitieren

Wendt, A., Possegger, H., Bittner, M., Schnöll, D., Wess, M., Malić, D., Bischof, H., & Jantsch, A. (2023). A Pedestrian Detection Case Study for a Traffic Light Controller. in S. Pasricha, & M. Shafique (Hrsg.), Embedded Machine Learning for Cyber-Physical, IoT, and Edge Computing: Software Optimazations and Hardware/Software Codesign (S. 75-96). Springer. https://doi.org/10.1007/978-3-031-39932-9_4

A Pedestrian Detection Case Study for a Traffic Light Controller. / Wendt, Alexander; Possegger, Horst; Bittner, Matthias et al.
Embedded Machine Learning for Cyber-Physical, IoT, and Edge Computing: Software Optimazations and Hardware/Software Codesign. Hrsg. / Sudeep Pasricha; Muhammad Shafique. Cham: Springer, 2023. S. 75-96.

Publikation: Beitrag in Buch/Bericht/Konferenzband › Beitrag in Buch/Bericht › Begutachtung

Wendt, A, Possegger, H, Bittner, M, Schnöll, D, Wess, M, Malić, D , Bischof, H & Jantsch, A 2023, A Pedestrian Detection Case Study for a Traffic Light Controller. in S Pasricha & M Shafique (Hrsg.), Embedded Machine Learning for Cyber-Physical, IoT, and Edge Computing: Software Optimazations and Hardware/Software Codesign. Springer, Cham, S. 75-96. https://doi.org/10.1007/978-3-031-39932-9_4

@inbook{9771caff68914d74b449ad81f4ab7a6d,

title = "A Pedestrian Detection Case Study for a Traffic Light Controller",

abstract = "A pedestrian detection system in a traffic light controller is studied. The system is based on Deep Neural Networks (DNNs). We explore several network architectures and hardware platforms to identify the most suitable solution under the given constraints of latency, cost, and precision. Specifically, we study altogether 13 networks from the MobileNet, Yolo, ResNet, and EfficientDet families and 6 platforms based on Nvidia and Intel platforms, conducting 383 experiments. We find that several network-platform combinations meet the given requirements of maximum 100 ms inference latency and 0.9 mean average precision. The most promising are Yolo v5 networks on Nvidia Jetson TX2 and IntelNUC GPU hardware.",

author = "Alexander Wendt and Horst Possegger and Matthias Bittner and Daniel Schn{\"o}ll and Matthias Wess and Du{\v s}an Mali{\'c} and Horst Bischof and Axel Jantsch",

year = "2023",

doi = "10.1007/978-3-031-39932-9_4",

language = "English",

isbn = "978-3-031-39931-2",

pages = "75--96",

editor = "Sudeep Pasricha and Muhammad Shafique",

booktitle = "Embedded Machine Learning for Cyber-Physical, IoT, and Edge Computing",

publisher = "Springer",

}

TY - CHAP

T1 - A Pedestrian Detection Case Study for a Traffic Light Controller

AU - Wendt, Alexander

AU - Possegger, Horst

AU - Bittner, Matthias

AU - Schnöll, Daniel

AU - Wess, Matthias

AU - Malić, Dušan

AU - Bischof, Horst

AU - Jantsch, Axel

PY - 2023

Y1 - 2023

N2 - A pedestrian detection system in a traffic light controller is studied. The system is based on Deep Neural Networks (DNNs). We explore several network architectures and hardware platforms to identify the most suitable solution under the given constraints of latency, cost, and precision. Specifically, we study altogether 13 networks from the MobileNet, Yolo, ResNet, and EfficientDet families and 6 platforms based on Nvidia and Intel platforms, conducting 383 experiments. We find that several network-platform combinations meet the given requirements of maximum 100 ms inference latency and 0.9 mean average precision. The most promising are Yolo v5 networks on Nvidia Jetson TX2 and IntelNUC GPU hardware.

AB - A pedestrian detection system in a traffic light controller is studied. The system is based on Deep Neural Networks (DNNs). We explore several network architectures and hardware platforms to identify the most suitable solution under the given constraints of latency, cost, and precision. Specifically, we study altogether 13 networks from the MobileNet, Yolo, ResNet, and EfficientDet families and 6 platforms based on Nvidia and Intel platforms, conducting 383 experiments. We find that several network-platform combinations meet the given requirements of maximum 100 ms inference latency and 0.9 mean average precision. The most promising are Yolo v5 networks on Nvidia Jetson TX2 and IntelNUC GPU hardware.

U2 - 10.1007/978-3-031-39932-9_4

DO - 10.1007/978-3-031-39932-9_4

M3 - Chapter

SN - 978-3-031-39931-2

SP - 75

EP - 96

BT - Embedded Machine Learning for Cyber-Physical, IoT, and Edge Computing

A2 - Pasricha, Sudeep

A2 - Shafique, Muhammad

PB - Springer

CY - Cham

ER -

A Pedestrian Detection Case Study for a Traffic Light Controller

Abstract

Zugriff auf Dokument

Fingerprint

Dieses zitieren