Car crash detection using ensemble deep learning and multimodal data from dashboard cameras

作者：

Highlights：

• A car crash detection system based on an ensemble deep learning model is proposed.

• An ensemble deep learning model uses video and audio data from dashboard cameras.

• A GRU-and-CNN-based base classifier is developed for video data.

• A GRU-based and a CNN-based base classifiers are developed for audio data.

• The proposed system establishes state-of-the-art classification performances.

摘要

•A car crash detection system based on an ensemble deep learning model is proposed.•An ensemble deep learning model uses video and audio data from dashboard cameras.•A GRU-and-CNN-based base classifier is developed for video data.•A GRU-based and a CNN-based base classifiers are developed for audio data.•The proposed system establishes state-of-the-art classification performances.

论文关键词：Dashboard camera,Car crash,Multimodal data,Deep learning,Ensemble technique

论文评审过程：Received 7 July 2020, Revised 7 June 2021, Accepted 8 June 2021, Available online 13 June 2021, Version of Record 19 June 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115400