license: other | |
task_categories: | |
- video-classification | |
language: | |
- en | |
tags: | |
- accident | |
- crash | |
- motorcycle | |
dataset_info: | |
features: | |
- name: filename | |
dtype: string | |
- name: case | |
dtype: string | |
- name: duration | |
dtype: float | |
# Dataset Summary | |
The dataset consisted of 2 types of cases; accident and driving while riding a motorcycle. 68 accident cases and 68 driving cases are prepared. 30 fps and 852x480 by default. It might be helpful when you train a model to infer whether a video is a motorcycle crash or not. One thing you should know about is 'driving videos' are not typically motorcycle driving. Most 'driving videos' are dashcams in the car. However, all the videos about accidents are motorcycle traffic accidents. | |