NORMA eResearch @NCI Library

Traffic Sign Detection and Recognition for Autonomous Vehicles Using Transfer Learning

Potla, Naresh (2023) Traffic Sign Detection and Recognition for Autonomous Vehicles Using Transfer Learning. Masters thesis, Dublin, National College of Ireland.

[thumbnail of Master of Science]
PDF (Master of Science)
Download (2MB) | Preview
[thumbnail of Configuration manual]
PDF (Configuration manual)
Download (1MB) | Preview


Driving a vehicle requires a number of important duties, one of which is identifying and recognising traffic signboards. They offer details regarding the state of the roads so that you can drive safely and enjoyably. The potential uses for traffic sign recognition in automated driving are numerous. Small traffic signs’ ability to be recognized will be impacted by external elements such as lighting, fading colours of indicator markers, climate, and diffraction. One of the main causes of car accidents is carelessness, specifically a failure to comprehend and misrepresenting traffic signs. A driving simulation program in-vehicle has the ability to give drivers new visual feedback for a better drive. This research demonstrates various transfer learning methods like EfficienNetV2L and VGG19 used for classifying the different signboards. To achieve a better performance model, I applied various data augmentation techniques with the help of Image Data Generator, which is available in Keras Library. Achieved an accuracy of 78% using VGG19, and EfficientV2L outperforms the VGG19 model with an improved accuracy of 93%.

Item Type: Thesis (Masters)
Anant, Aaloka
Subjects: Q Science > QA Mathematics > Electronic computers. Computer science
T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science
T Technology > TE Highway engineering. Roads and pavements
T Technology > TL Motor vehicles. Aeronautics. Astronautics
Divisions: School of Computing > Master of Science in Data Analytics
Depositing User: Tamara Malone
Date Deposited: 24 May 2023 17:30
Last Modified: 24 May 2023 17:30

Actions (login required)

View Item View Item