Scholarlink

Driver gaze estimation is important for different driving applications such as driver gaze behavior understanding, visual distraction detection, and making gaze-based advanced driving assistance systems (ADAS). This study aims to build a driver gaze estimation model based on gaze zone classification using a deep neural network. The driver’s visual field is divided into pre-defined gaze zones, where the driver frequently looks during driving. A benchmark driver gaze dataset named Driver Gaze in Wild (DGW) was used for this study. A pre-trained convolutional neural network (CNN) model, EfficentNet-B7 model was fine tuning for the driver gaze zone classification. A classification report and confusion matrix were plotted to show the result of model training and performance. Finally, we checked the gaze classification model on the driver face video driving data to track the driver gaze.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Driver Gaze Zone Estimation Using Deep Neural Network

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Pavan Kumar Sharma,
  • Pranamesh Chakraborty

摘要

Driver gaze estimation is important for different driving applications such as driver gaze behavior understanding, visual distraction detection, and making gaze-based advanced driving assistance systems (ADAS). This study aims to build a driver gaze estimation model based on gaze zone classification using a deep neural network. The driver’s visual field is divided into pre-defined gaze zones, where the driver frequently looks during driving. A benchmark driver gaze dataset named Driver Gaze in Wild (DGW) was used for this study. A pre-trained convolutional neural network (CNN) model, EfficentNet-B7 model was fine tuning for the driver gaze zone classification. A classification report and confusion matrix were plotted to show the result of model training and performance. Finally, we checked the gaze classification model on the driver face video driving data to track the driver gaze.