Publication Date
12-2019
Document Type
Thesis
Keywords
autonomous vehicle, driverless vehicle, robot vehicle, self-driving, convolutional neural network, ConvNet model, Alexnet
Abstract
The risk of pedestrian accidents has increased due to the distracted walking increase. The research in the autonomous vehicles industry aims to minimize this risk by enhancing the route planning to produce safer routes. Detecting distracted pedestrians plays a significant role in identifying safer routes and hence decreases pedestrian accident risk. Thus, this research aims to investigate how to use the convolutional neural networks for building an algorithm that significantly improves the accuracy of detecting distracted pedestrians based on gathered cues. Particularly, this research involves the analysis of pedestrian’ images to identify distracted pedestrians who are not paying attention when crossing the road. This work tested three different architectures of convolutional neural networks. These architectures are Basic, Deep, and AlexNet. The performance of the three architectures was evaluated based on two datasets. The first is a new training dataset called SCIT and created by this work based on recorded videos of volunteers from Sheridan College Institute of Technology. The second is a public dataset called PETA, which was made up of images with various resolutions. The ConvNet model with the Deep architecture outperformed the Basic and AlexNet architectures in detecting distracted pedestrians.
Faculty
Faculty of Applied Science & Technology
Copyright
© Igor Grishchenko
Terms of Use
Terms of Use for Works posted in SOURCE.
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.
Original Publication Citation
Grishchenko, I. (2019). Detection of distracted pedestrians using convolutional neural networks (Unpublished thesis). Sheridan College, Ontario, Canada.
SOURCE Citation
Grishchenko, Igor, "Detection of Distracted Pedestrians using Convolutional Neural Networks" (2019). Faculty of Applied Science and Technology - Exceptional Student Work, Applied Computing Theses. 7.
https://source.sheridancollege.ca/student_work_fast_applied_computing_theses/7
