CAMDet: CAM-based objection detection for non-crowded views from moving IoT devices

Yuheng Cao; Kwei Jay Lin; Bo Lung Tsai; Yu Meng

doi:10.1109/iThings-GreenCom-CPSCom-SmartData-Cybermatics50389.2020.00093

CAMDet: CAM-based objection detection for non-crowded views from moving IoT devices

Yuheng Cao, Kwei Jay Lin, Bo Lung Tsai, Yu Meng

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

The integration of Internet of Things (IoT) and Artificial Intelligence (AI) brings us AIoT that delivers the capabilities of object detection, device localization, object tracking and re-identification on moving IoT devices to control robots/drones for smart human-machine interactions. Among the tasks, efficient objection detection plays an importance role since it acts as the foundation of many other vision-based IoT applications. One main challenge is to locate target objects fast and accurate. The paper presents the CAMDet technology that utilizes Class Activation Map (CAM) to reduce the convolution blocks and the enormous candidate bounding boxes in the detection-head stage. We have designed CAMDet and integrate it with other backbone networks. CAMDet is shown to be 2.1-2.7 times faster than the popular Tiny-YOLO/SSD methods in non-crowded scenarios when using the same backbone and feature pyramid structure. The performance study shows that our proposed methods are very attractive for real time object detection on moving IoT devices.

Original language	English
Title of host publication	Proceedings - IEEE Congress on Cybermatics
Subtitle of host publication	2020 IEEE International Conferences on Internet of Things, iThings 2020, IEEE Green Computing and Communications, GreenCom 2020, IEEE Cyber, Physical and Social Computing, CPSCom 2020 and IEEE Smart Data, SmartData 2020
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	519-526
Number of pages	8
ISBN (Electronic)	9781728176475
DOIs	https://doi.org/10.1109/iThings-GreenCom-CPSCom-SmartData-Cybermatics50389.2020.00093
State	Published - 11 2020
Externally published	Yes
Event	2020 IEEE Congress on Cybermatics: 13th IEEE International Conferences on Internet of Things, iThings 2020, 16th IEEE International Conference on Green Computing and Communications, GreenCom 2020, 13th IEEE International Conference on Cyber, Physical and Social Computing, CPSCom 2020 and 6th IEEE International Conference on Smart Data, SmartData 2020 - Rhodes Island, Greece Duration: 02 11 2020 → 06 11 2020

Publication series

Name	Proceedings - IEEE Congress on Cybermatics: 2020 IEEE International Conferences on Internet of Things, iThings 2020, IEEE Green Computing and Communications, GreenCom 2020, IEEE Cyber, Physical and Social Computing, CPSCom 2020 and IEEE Smart Data, SmartData 2020

Conference

Conference	2020 IEEE Congress on Cybermatics: 13th IEEE International Conferences on Internet of Things, iThings 2020, 16th IEEE International Conference on Green Computing and Communications, GreenCom 2020, 13th IEEE International Conference on Cyber, Physical and Social Computing, CPSCom 2020 and 6th IEEE International Conference on Smart Data, SmartData 2020
Country/Territory	Greece
City	Rhodes Island
Period	02/11/20 → 06/11/20

Bibliographical note

Publisher Copyright:
© 2020 IEEE.

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1109/iThings-GreenCom-CPSCom-SmartData-Cybermatics50389.2020.00093

Cite this

Cao, Y., Lin, K. J., Tsai, B. L., & Meng, Y. (2020). CAMDet: CAM-based objection detection for non-crowded views from moving IoT devices. In Proceedings - IEEE Congress on Cybermatics: 2020 IEEE International Conferences on Internet of Things, iThings 2020, IEEE Green Computing and Communications, GreenCom 2020, IEEE Cyber, Physical and Social Computing, CPSCom 2020 and IEEE Smart Data, SmartData 2020 (pp. 519-526). Article 9291561 (Proceedings - IEEE Congress on Cybermatics: 2020 IEEE International Conferences on Internet of Things, iThings 2020, IEEE Green Computing and Communications, GreenCom 2020, IEEE Cyber, Physical and Social Computing, CPSCom 2020 and IEEE Smart Data, SmartData 2020). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/iThings-GreenCom-CPSCom-SmartData-Cybermatics50389.2020.00093

Cao, Yuheng ; Lin, Kwei Jay ; Tsai, Bo Lung et al. / CAMDet : CAM-based objection detection for non-crowded views from moving IoT devices. Proceedings - IEEE Congress on Cybermatics: 2020 IEEE International Conferences on Internet of Things, iThings 2020, IEEE Green Computing and Communications, GreenCom 2020, IEEE Cyber, Physical and Social Computing, CPSCom 2020 and IEEE Smart Data, SmartData 2020. Institute of Electrical and Electronics Engineers Inc., 2020. pp. 519-526 (Proceedings - IEEE Congress on Cybermatics: 2020 IEEE International Conferences on Internet of Things, iThings 2020, IEEE Green Computing and Communications, GreenCom 2020, IEEE Cyber, Physical and Social Computing, CPSCom 2020 and IEEE Smart Data, SmartData 2020).

@inproceedings{e4d328a7710d4ecf88712daff6ce8f40,

title = "CAMDet: CAM-based objection detection for non-crowded views from moving IoT devices",

abstract = "The integration of Internet of Things (IoT) and Artificial Intelligence (AI) brings us AIoT that delivers the capabilities of object detection, device localization, object tracking and re-identification on moving IoT devices to control robots/drones for smart human-machine interactions. Among the tasks, efficient objection detection plays an importance role since it acts as the foundation of many other vision-based IoT applications. One main challenge is to locate target objects fast and accurate. The paper presents the CAMDet technology that utilizes Class Activation Map (CAM) to reduce the convolution blocks and the enormous candidate bounding boxes in the detection-head stage. We have designed CAMDet and integrate it with other backbone networks. CAMDet is shown to be 2.1-2.7 times faster than the popular Tiny-YOLO/SSD methods in non-crowded scenarios when using the same backbone and feature pyramid structure. The performance study shows that our proposed methods are very attractive for real time object detection on moving IoT devices.",

author = "Yuheng Cao and Lin, {Kwei Jay} and Tsai, {Bo Lung} and Yu Meng",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE.; 2020 IEEE Congress on Cybermatics: 13th IEEE International Conferences on Internet of Things, iThings 2020, 16th IEEE International Conference on Green Computing and Communications, GreenCom 2020, 13th IEEE International Conference on Cyber, Physical and Social Computing, CPSCom 2020 and 6th IEEE International Conference on Smart Data, SmartData 2020 ; Conference date: 02-11-2020 Through 06-11-2020",

year = "2020",

month = nov,

doi = "10.1109/iThings-GreenCom-CPSCom-SmartData-Cybermatics50389.2020.00093",

language = "英语",

series = "Proceedings - IEEE Congress on Cybermatics: 2020 IEEE International Conferences on Internet of Things, iThings 2020, IEEE Green Computing and Communications, GreenCom 2020, IEEE Cyber, Physical and Social Computing, CPSCom 2020 and IEEE Smart Data, SmartData 2020",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "519--526",

booktitle = "Proceedings - IEEE Congress on Cybermatics",

address = "美国",

}

Cao, Y, Lin, KJ, Tsai, BL & Meng, Y 2020, CAMDet: CAM-based objection detection for non-crowded views from moving IoT devices. in Proceedings - IEEE Congress on Cybermatics: 2020 IEEE International Conferences on Internet of Things, iThings 2020, IEEE Green Computing and Communications, GreenCom 2020, IEEE Cyber, Physical and Social Computing, CPSCom 2020 and IEEE Smart Data, SmartData 2020., 9291561, Proceedings - IEEE Congress on Cybermatics: 2020 IEEE International Conferences on Internet of Things, iThings 2020, IEEE Green Computing and Communications, GreenCom 2020, IEEE Cyber, Physical and Social Computing, CPSCom 2020 and IEEE Smart Data, SmartData 2020, Institute of Electrical and Electronics Engineers Inc., pp. 519-526, 2020 IEEE Congress on Cybermatics: 13th IEEE International Conferences on Internet of Things, iThings 2020, 16th IEEE International Conference on Green Computing and Communications, GreenCom 2020, 13th IEEE International Conference on Cyber, Physical and Social Computing, CPSCom 2020 and 6th IEEE International Conference on Smart Data, SmartData 2020, Rhodes Island, Greece, 02/11/20. https://doi.org/10.1109/iThings-GreenCom-CPSCom-SmartData-Cybermatics50389.2020.00093

CAMDet: CAM-based objection detection for non-crowded views from moving IoT devices. / Cao, Yuheng; Lin, Kwei Jay; Tsai, Bo Lung et al.
Proceedings - IEEE Congress on Cybermatics: 2020 IEEE International Conferences on Internet of Things, iThings 2020, IEEE Green Computing and Communications, GreenCom 2020, IEEE Cyber, Physical and Social Computing, CPSCom 2020 and IEEE Smart Data, SmartData 2020. Institute of Electrical and Electronics Engineers Inc., 2020. p. 519-526 9291561 (Proceedings - IEEE Congress on Cybermatics: 2020 IEEE International Conferences on Internet of Things, iThings 2020, IEEE Green Computing and Communications, GreenCom 2020, IEEE Cyber, Physical and Social Computing, CPSCom 2020 and IEEE Smart Data, SmartData 2020).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - CAMDet

T2 - 2020 IEEE Congress on Cybermatics: 13th IEEE International Conferences on Internet of Things, iThings 2020, 16th IEEE International Conference on Green Computing and Communications, GreenCom 2020, 13th IEEE International Conference on Cyber, Physical and Social Computing, CPSCom 2020 and 6th IEEE International Conference on Smart Data, SmartData 2020

AU - Cao, Yuheng

AU - Lin, Kwei Jay

AU - Tsai, Bo Lung

AU - Meng, Yu

PY - 2020/11

Y1 - 2020/11

N2 - The integration of Internet of Things (IoT) and Artificial Intelligence (AI) brings us AIoT that delivers the capabilities of object detection, device localization, object tracking and re-identification on moving IoT devices to control robots/drones for smart human-machine interactions. Among the tasks, efficient objection detection plays an importance role since it acts as the foundation of many other vision-based IoT applications. One main challenge is to locate target objects fast and accurate. The paper presents the CAMDet technology that utilizes Class Activation Map (CAM) to reduce the convolution blocks and the enormous candidate bounding boxes in the detection-head stage. We have designed CAMDet and integrate it with other backbone networks. CAMDet is shown to be 2.1-2.7 times faster than the popular Tiny-YOLO/SSD methods in non-crowded scenarios when using the same backbone and feature pyramid structure. The performance study shows that our proposed methods are very attractive for real time object detection on moving IoT devices.

AB - The integration of Internet of Things (IoT) and Artificial Intelligence (AI) brings us AIoT that delivers the capabilities of object detection, device localization, object tracking and re-identification on moving IoT devices to control robots/drones for smart human-machine interactions. Among the tasks, efficient objection detection plays an importance role since it acts as the foundation of many other vision-based IoT applications. One main challenge is to locate target objects fast and accurate. The paper presents the CAMDet technology that utilizes Class Activation Map (CAM) to reduce the convolution blocks and the enormous candidate bounding boxes in the detection-head stage. We have designed CAMDet and integrate it with other backbone networks. CAMDet is shown to be 2.1-2.7 times faster than the popular Tiny-YOLO/SSD methods in non-crowded scenarios when using the same backbone and feature pyramid structure. The performance study shows that our proposed methods are very attractive for real time object detection on moving IoT devices.

UR - http://www.scopus.com/inward/record.url?scp=85099440493&partnerID=8YFLogxK

U2 - 10.1109/iThings-GreenCom-CPSCom-SmartData-Cybermatics50389.2020.00093

DO - 10.1109/iThings-GreenCom-CPSCom-SmartData-Cybermatics50389.2020.00093

M3 - 会议稿件

AN - SCOPUS:85099440493

T3 - Proceedings - IEEE Congress on Cybermatics: 2020 IEEE International Conferences on Internet of Things, iThings 2020, IEEE Green Computing and Communications, GreenCom 2020, IEEE Cyber, Physical and Social Computing, CPSCom 2020 and IEEE Smart Data, SmartData 2020

SP - 519

EP - 526

BT - Proceedings - IEEE Congress on Cybermatics

PB - Institute of Electrical and Electronics Engineers Inc.

Y2 - 2 November 2020 through 6 November 2020

ER -

Cao Y, Lin KJ, Tsai BL, Meng Y. CAMDet: CAM-based objection detection for non-crowded views from moving IoT devices. In Proceedings - IEEE Congress on Cybermatics: 2020 IEEE International Conferences on Internet of Things, iThings 2020, IEEE Green Computing and Communications, GreenCom 2020, IEEE Cyber, Physical and Social Computing, CPSCom 2020 and IEEE Smart Data, SmartData 2020. Institute of Electrical and Electronics Engineers Inc. 2020. p. 519-526. 9291561. (Proceedings - IEEE Congress on Cybermatics: 2020 IEEE International Conferences on Internet of Things, iThings 2020, IEEE Green Computing and Communications, GreenCom 2020, IEEE Cyber, Physical and Social Computing, CPSCom 2020 and IEEE Smart Data, SmartData 2020). doi: 10.1109/iThings-GreenCom-CPSCom-SmartData-Cybermatics50389.2020.00093

CAMDet: CAM-based objection detection for non-crowded views from moving IoT devices

Abstract

Publication series

Conference

Bibliographical note

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this