Autonomous vehicle design is the present-day research field in which traffic sign recognition and classification is the challenging task. The motive of this research work is to minimize the involvement of human consciousness. Automatic traffic sign recognition systems are designed so that vehicles are capable of choosing right direction, speed, etc. This research paper focuses on the sign classification function. Here we proposed a deep learning approach to design a model to classify road traffic signs. Convolutional neural network (CNN) architecture is designed for the German Traffic Sign Recognition Benchmark (GTSRB) dataset which passes through four layers of convolution operation with increasing filter size and two fully connected layers. Data augmentation and batch normalization techniques are the highlights of this research. Dropout is used to prevent over-fitting and this model achieved 98.65% accuracy in predicting the exact class of unseen images.
Deep Learning Approach to Classify Road Traffic Sign Images
Lect. Notes in Networks, Syst.
International Conference on Image Processing and Capsule Networks ; 2021 ; Bangkok, Thailand May 27, 2021 - May 28, 2021
Second International Conference on Image Processing and Capsule Networks ; Kapitel : 14 ; 146-155
10.09.2021
10 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Road Safety with Deep Learning Voice Based Traffic Sign
IEEE | 2025
|A Deep-Learning Approach to Detect and Classify Heavy-Duty Trucks in Satellite Images
IEEE | 2024
|Traffic Sign Detection and Recognition Using Deep Learning Approach
Springer Verlag | 2023
|Traffic Sign Classification using Deep Learning
IEEE | 2023
|