Abstract:
With the development of artificial intelligence, autonomous vehicles technology has also achieved breakthrough results, and these vehicles have also begun to be populariz...Show MoreMetadata
Abstract:
With the development of artificial intelligence, autonomous vehicles technology has also achieved breakthrough results, and these vehicles have also begun to be popularized in the market. However, with the growing prevalence of autonomous vehicles, there has been a corresponding increase in traffic accidents attributed to distracted driving. This rise is partly due to drivers overtrusting self-driving systems that have not yet achieved full autonomy, resulting in increased driver distraction. Consequently, detecting drivers’ distracted behaviors remains a critical necessity. To address this issue, this study introduces a spatial-aware feature aggregation network (SFANet) for driver distraction detection (DDD). This method utilizes a spatial-aware feature aggregation attention mechanism (SFAAM) to capture a long range of spatially aware features. Experimental results indicate that the SFANet model demonstrates superior performance compared to previous studies on the State Farm Distracted Driver Detection (SFD3) database, achieving an impressive accuracy of 99.84% on the test dataset.
Date of Conference: 09-11 July 2024
Date Added to IEEE Xplore: 18 September 2024
ISBN Information: