Program Type
Honors
Faculty Advisor
Kaiman Zeng
Document Type
Presentation
Location
Face-to-face
Start Date
25-4-2023 11:24 AM
Abstract
As the world continues to become more technologically advanced, distracted driving will continue to be a growing danger to the public. Convolutional neural networks can be used to monitor driving and differentiate distracted driving from safe driving. A popular distracted driving dataset created by State Farm called the Distracted Driver Dataset can be trained with the Auto-Keras model API. Auto-Keras is a system that taylors a machine learning model to fit a given dataset. While experienced neural network designers can create neural networks to produce incredibly accurate results, Auto-Keras gives those with less expertise a method of designing a network in a time-efficient manner. The resulting network created through Auto-Keras achieved an accuracy of 95.5%
Recommended Citation
Heikes, Wesley M., "A Study of Deep Neural Networks in the Application of Distracted Driving Detection" (2023). ATU Research Symposium. 51.
https://orc.library.atu.edu/atu_rs/2023/2023/51
Included in
A Study of Deep Neural Networks in the Application of Distracted Driving Detection
Face-to-face
As the world continues to become more technologically advanced, distracted driving will continue to be a growing danger to the public. Convolutional neural networks can be used to monitor driving and differentiate distracted driving from safe driving. A popular distracted driving dataset created by State Farm called the Distracted Driver Dataset can be trained with the Auto-Keras model API. Auto-Keras is a system that taylors a machine learning model to fit a given dataset. While experienced neural network designers can create neural networks to produce incredibly accurate results, Auto-Keras gives those with less expertise a method of designing a network in a time-efficient manner. The resulting network created through Auto-Keras achieved an accuracy of 95.5%