Please use this identifier to cite or link to this item: http://hdl.handle.net/10316/94057
Title: Development of a Highway Tolling and Enforcement System Using Convolutional Neural Networks and Fine-Grained Visual Classification
Other Titles: Development of a Highway Tolling and Enforcement System Using Convolutional Neural Networks and Fine-Grained Visual Classification
Authors: Oliveira, Roberto Manuel Trindade
Orientador: Batista, Jorge Manuel Moreira de Campos Pereira
Keywords: Deep Learning; Convolutional Neural Networks; Fine-Grained Visual Classification; Pattern Recognition; Computer Vision; Deep Learning; Convolutional Neural Networks; Fine-Grained Visual Classification; Pattern Recognition; Computer Vision
Issue Date: 13-Nov-2020
Serial title, monograph or event: Development of a Highway Tolling and Enforcement System Using Convolutional Neural Networks and Fine-Grained Visual Classification
Place of publication or event: DEEC
Abstract: Deep Learning and Convolutional Neural Networks have been staples in solving challenges relatedto Image Processing, Computer Vision and Pattern Recognition. Since their breakthrough in 2012 thatno other method has come close, be it in overall results, consistency, but also computation capabilities,with the technology evolving and delivering more and more impressive results as framework developerspush the boundaries. As a consequence, we are seeing work in the domain of Deep Learning creepmore and more in our ever day lives, automating a variety of tasks.In this work, Deep Learning will be applied to try and automate one such barely noticed task: tollcollection. In Porugal, Brisa operates an automatic toll collection service, which despite their bestefforts, is still fragile and subject to fraud. It is then in their interest that toll collection becomesas precise and reliable as possible. With this in mind, this document explores technology relatedto Vehicle Recognition, and applies state-of-the-art methodologies that ultimately deliver a solutionthat performs with state-of-the-art competitive results. The methodologies here applied can easily bereplicated and should translate well to other critical aspects of social life, like medical imaging baseddiagnosing.This work was submitted for the partial fulfillment of the requirements to complete the Integrated Master in Electrical and Computer Engineering, Automation Specialization, on October 2020. The work was supervised by Professor Jorge Manuel Moreira de Campos Pereira Batista, Phd.
Deep Learning and Convolutional Neural Networks have been staples in solving challenges relatedto Image Processing, Computer Vision and Pattern Recognition. Since their breakthrough in 2012 thatno other method has come close, be it in overall results, consistency, but also computation capabilities,with the technology evolving and delivering more and more impressive results as framework developerspush the boundaries. As a consequence, we are seeing work in the domain of Deep Learning creepmore and more in our ever day lives, automating a variety of tasks.In this work, Deep Learning will be applied to try and automate one such barely noticed task: tollcollection. In Porugal, Brisa operates an automatic toll collection service, which despite their bestefforts, is still fragile and subject to fraud. It is then in their interest that toll collection becomesas precise and reliable as possible. With this in mind, this document explores technology relatedto Vehicle Recognition, and applies state-of-the-art methodologies that ultimately deliver a solutionthat performs with state-of-the-art competitive results. The methodologies here applied can easily bereplicated and should translate well to other critical aspects of social life, like medical imaging baseddiagnosing.This work was submitted for the partial fulfillment of the requirements to complete the Integrated Master in Electrical and Computer Engineering, Automation Specialization, on October 2020. The work was supervised by Professor Jorge Manuel Moreira de Campos Pereira Batista, Phd.
Description: Dissertação de Mestrado Integrado em Engenharia Electrotécnica e de Computadores apresentada à Faculdade de Ciências e Tecnologia
URI: http://hdl.handle.net/10316/94057
Rights: openAccess
Appears in Collections:UC - Dissertações de Mestrado

Files in This Item:
File Description SizeFormat
Dissertacao_Roberto_Oliveira_v2_final-1.pdf14.94 MBAdobe PDFView/Open
Show full item record

Page view(s)

21
checked on Jul 22, 2021

Download(s)

46
checked on Jul 22, 2021

Google ScholarTM

Check


This item is licensed under a Creative Commons License Creative Commons