Search

Due to the increase in motor vehicle accidents, there is a growing need for high-performance car crash detection systems. The authors of this research propose a car crash detection system that uses both video data and audio data from dashboard cameras in order to improve car crash detection performance. While most existing car crash detection systems depend on single modal data (i.e., video data or audio data only), the proposed car crash detection system uses an ensemble deep learning model based on multimodal data (i.e., both video and audio data), because different types of data extracted from one information source (e.g., dashboard cameras) can be regarded as different views of the same source. These different views complement one another and improve detection performance, because one view may have information that the other view does not contain. In this research, deep learning techniques, gated recurrent unit (GRU) and convolutional neural network (CNN), are used to develop a car crash detection system. A weighted average ensemble is used as an ensemble technique. The proposed car crash detection system, which is based on multiple classifiers that use both video and audio data from dashboard cameras, is validated using a comparison with single classifiers that use video data or audio data only. Car accident YouTube clips are used to validate this research. The experimental results indicate that the proposed car crash detection system performs significantly better than single classifiers. It is expected that the proposed car crash detection system can be used as part of an emergency road call service that recognizes traffic accidents automatically and allows immediate rescue after transmission to emergency recovery agencies.

Spectroscopic and microscopic techniques are not suitable for the rapid monitoring of time-dependent dissolution behavior (particularly, pinhole changes) of a layer deposited by atomic layer deposition (ALD). Here, we present a simple electrochemical method that provides information on the dissolution mechanism including pinhole generation and thickness change. Because indium tin oxide (ITO) electrodes exhibit flat capacitive currents and good (electro)chemical stability, they are selected as ideal underlying substrates for the electrochemical monitoring of the ALD layers even under harsh conditions. Two ALD layers (Al2O3 and TiO2 layers) that exhibit opposite dissolution behaviors are chosen as model layers because the as-deposited Al2O3 layers are pinhole-free but unstable in aqueous solutions, whereas the as-deposited TiO2 layers are not pinhole-free but stable in aqueous solutions. The combination of capacitive current level (in an electrolyte solution containing no redox-active species) and electrochemical blocking behavior (in an electrolyte solution containing a redox-active species such as Ru(NH3)63+ and ferrocenemethanol) obtained from cyclic voltammograms enables us to verify whether the dissolution of an ALD layer occurs, to evaluate the dissolution rate, and to identify the plausible dissolution mechanism. The electrochemical results reveal that the Al2O3 layers are dissolved in biological buffers, along with pinhole generation, and that the TiO2 layers are stable with no pinhole generation. The difference in electrochemical blocking behavior between Ru(NH3)63+ and ferrocenemethanol provides information on the approximate size of the pinholes. The present method is appealing for practical use because even an ALD layer with a thickness of only a few nanometers can be tested to monitor the dissolution behavior and because any ALD layer that can be readily deposited on ITO electrodes can be easily examined using this method.