Presentation 2023-10-19
An onboard vehicle camera object tracking for data labeling of object detection by Federated Learning
Yuki Nakahama, Satoshi Ohzahata, Ryo Yamamoto,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In recent years, auto vehicle driving technology has made significant advancements. An achieving fully automatic driving requires accurate information acquisition about objects around the vehicle. Consequently, a substantial amount of training data is necessary. In the labelling process, since it is difficult to label manually, object detection models are often used. However, the labeling accuracy for distant objects in existing object detection models is still low and this affects quality of the training. In addition, the training process of Federated Learning needs to be conducted locally due to privacy concern. This paper proposes to improve labeling precision by employing Multi-Object Tracking (MoT) with multiple cameras to apply the results of nearby objects to identify distant ones in time-series data.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Object Tracking / Labeling / Federated Learning
Paper # IN2023-33
Date of Issue 2023-10-12 (IN)

Conference Information
Committee IN / ICTSSL / IEE-SMF
Conference Date 2023/10/19(2days)
Place (in Japanese) (See Japanese page)
Place (in English) Fukuoka University
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair Kunio Hato(NTT) / Osamu Uchida(Tokai Univ.) / 石田 隆張(明星大学)
Vice Chair Tsutomu Murase(Nagoya Univ.) / Munenari Inoguchi(Toyama Univ.) / Akira Otsuka(Graduate Institute for Entrepreneurial Studies)
Secretary Tsutomu Murase(NTT) / Munenari Inoguchi(NTT) / Akira Otsuka(KDDI Research) / (Fukuoka Univ.)
Assistant / Makoto Kobayashi(Hiroshima City Univ.)

Paper Information
Registration To Technical Committee on Information Networks / Technical Committee on Information and Communication Technologies for Safe and Secure Life / Technical Committee on Smart Facilities
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) An onboard vehicle camera object tracking for data labeling of object detection by Federated Learning
Sub Title (in English)
Keyword(1) Object Tracking
Keyword(2) Labeling
Keyword(3) Federated Learning
1st Author's Name Yuki Nakahama
1st Author's Affiliation The University of Electro-Communications(UEC)
2nd Author's Name Satoshi Ohzahata
2nd Author's Affiliation The University of Electro-Communications(UEC)
3rd Author's Name Ryo Yamamoto
3rd Author's Affiliation The University of Electro-Communications(UEC)
Date 2023-10-19
Paper # IN2023-33
Volume (vol) vol.123
Number (no) IN-219
Page pp.pp.5-10(IN),
#Pages 6
Date of Issue 2023-10-12 (IN)