Authors: Shouhei Kunitomi and Shinichi Takayama—Japan Automobile Research Institute; Masayuki Shirakawa—Japan Automobile Manufacturers Association, Inc.
Abstract
The goal of this study is to clarify the usefulness of deep learning methods for pedestrian collision detection using dashcam videos for advanced automatic collision notification, focusing on pedestrians, as they make up the highest number of traffic fatalities in Japan. First, we created a dataset for deep learning from dashcam videos. A total of 78 dashcam videos of pedestrian-to-automobile accidents were collected from a video hosting website and from the Japan Automobile Research Institute (JARI). Individual frames were selected from the video data amounting to a total of 1,212 still images, which were added to our dataset with class and location information. This dataset was then divided to create training, validation, and test datasets. Next, deep learning was performed based on the training dataset to learn the features of pedestrian collision images, which are images that capture pedestrian behavior at the time of the collision. Pedestrian collision detection performance of the trained model was evaluated as the percentage of correct predictions of pedestrian collisions in image data according to varied test sets with different combinations of characteristics. Our results for the proposed method show high-precision collision detection for daytime, clear pedestrian wrap trajectory accident data, including accurate detection of pedestrian collision location information. However, nighttime, unclear accident data resulted in false detection or no detection. Reduction of exposure value and resolution was confirmed to reduce detection rate. The results of the present study suggest the possibility of pedestrian collision detection by deep learning using dashcam videos.
Type: Full Paper
Keywords: AACN, collision detection, pedestrian, deep learning
© Stapp Association, 2020
Access Additional Papers from This Volume
View additional Full Papers from the Stapp Car Crash Journal, Volume 64.
- Analysis of Kinematic Response of Pediatric Occupants Seated in Naturalistic Positions in Simulated Frontal Small Offset Impacts: With and Without Automatic Emergency BrakingAuthors: Jalaj Maheshwari, Shreyas Sarfare, and Aditya Belwadi—Center for Injury Research and Prevention, Children’s Hospital of Philadelphia; Clayton Falciani—Center for…
- Application of Deep Learning Methods for Pedestrian Collision Detection Using Dashcam VideosAuthors: Shouhei Kunitomi and Shinichi Takayama—Japan Automobile Research Institute; Masayuki Shirakawa—Japan Automobile Manufacturers Association, Inc. Abstract The goal of this…
- Biomechanical Responses and Injury Assessment of Post Mortem Human Subjects in Various Rear-facing Seating ConfigurationsAuthors: Yun-Seok Kang—Injury Biomechanics Research Center, The Ohio State University; Jason Stammen and Kevin Moorhouse—National Highway Traffic Safety Administration, Vehicle…
- Effects of Technology on Drivers’ Behavior during Backing ManeuversAuthors: Yasuhiro Matsui—National Traffic Safety and Environment Laboratory, Japan; Shoko Oikawa—Tokyo Metropolitan University, Japan Abstract This paper examines how vehicle…
- Evaluation of Rotation Reduction Features in Infant and Extended-Use Convertible Child Restraint Systems during Frontal and Rear ImpactsAuthors: Declan A. Patton, Aditya N. Belwadi, and Jalaj Maheshwari—Center for Injury Research and Prevention, Children’s Hospital of Philadelphia; Kristy…
- Geometrical and Mechanical Characterization of the Abdominal Fold of Obese Post Mortem Human Subjects for Use in Human Body ModellingAuthors: Matthieu Lebarbé—CEESAR (France); Philippe Beillas, Tomas Janak, and Yoann Lafon—Univ Lyon, Univ Claude Bernard Lyon 1, Univ Gustave Eiffel,…
- Kinematic and Injury Response of Reclined PMHS in Frontal ImpactsAuthors: Rachel Richardson, John-Paul Donlon, Mohan Jayathirtha, Jason L. Forman, Greg Shaw, Bronislaw Gepner, and Jason R. Kerrigan—University of Virginia…
- Variations in User Implementation of the CORA Rating MetricAuthor: Devon L. Albert — Center for Injury Biomechanics, Department of Biomedical Engineering and Mechanics, Virginia Tech Abstract The CORA…