Search Results

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.

The objective of the current study was to provide a comprehensive characterization of human biomechanical response to whole-body, lateral impact. Three approximately 50th-percentile adult male PMHS were subjected to right-side pure lateral impacts at 4.3 ± 0.1 m/s using a rigid wall mounted to a rail-mounted sled. Each subject was positioned on a rigid seat and held stationary by a system of tethers until immediately prior to being impacted by the moving wall with 100 mm pelvic offset. Displacement data were obtained using an optoelectronic stereophotogrammetric system that was used to track the 3D motions of the impacting wall sled; seat sled, and reflective targets secured to the head, spine, extremities, ribcage, and shoulder complex of each subject. Kinematic data were also recorded using 3-axis accelerometer cubes secured to the head, pelvis, and spine at the levels of T1, T6, T11, and L3. Chest deformation in the transverse plane was recorded using a single chestband.