The 8th International Conference on Next Generation Computing 2022 (2022.10)바로가기
페이지
pp.256-258
저자
Ji-Hyeon Ryu, Hyung-Won Kim
언어
영어(ENG)
URL
https://www.earticle.net/Article/A419791
원문정보
초록
영어
Recently, as e-scooter rental services are growing, e-scooter is considered as one of the most convenient and inexpensive means of transportation. As the use of e-scooters has soared in recent years, traffic accidents between vehicles and e-scooters riders have also increased. Therefore, there is also a need for the development of deep learning techniques for detecting e-scooters and riders. However, there is no large dataset labeled for e-scooters on the road captured by dash cameras on vehicles. In this paper, a new e-scooter dataset has been constructed using vehicles’ dashcam views to address this problem. This dataset consists of running e-scooters, parked e-scooters, e-scooter riders, and pedestrians. This paper also presents a CNN model for detecting e-scooters and riders, and analyzes the detection performance using our e-scooter dataset.
목차
Abstract I. INTRODUCTION II. CONSTRUCTING DATASET III. TRAINING AND EVALUATION IV. CONCLUSION REFERENCES
키워드
e-scooterkickboarddetectionCNN
저자
Ji-Hyeon Ryu [ College of Electrical and Computer Engineering Chungbuk National University ]
Hyung-Won Kim [ College of Electrical and Computer Engineering Chungbuk National University ]
Corresponding Author