卷积神经网络
计算机科学
人工智能
特征(语言学)
面子(社会学概念)
特征提取
模式识别(心理学)
计算机视觉
社会科学
哲学
语言学
社会学
标识
DOI:10.1109/iscsic57216.2022.00040
摘要
Aiming at the problem of low accuracy of fatigue driving judgment based on single feature, a fatigue driving detection method based on multiple features was proposed. In order to improve the accuracy of fatigue driving judgment, the RetinaFace algorithm was first used to detect the driver’s face, and then the facial feature points were located to complete the estimation of head posture. Then, the end-to-end convolutional neural network model was used to identify the face state of the eyes and mouth. Finally, combined with the threshold of characteristic parameters of eyes, mouth and head, a fatigue driving detection method based on multiple features was proposed. The experimental results showed that the recognition accuracy of the improved convolutional neural network model for eyes and mouth is more than 95%. The judgment accuracy of the fatigue driving detection method based on feature fusion is more than 90%. Therefore, the fatigue driving detection method based on multiple features is feasible in fatigue driving judgment.
科研通智能强力驱动
Strongly Powered by AbleSci AI