Enhancing traffic sign recognition (TSR) by classifying deep learning models to promote road safety

• Published in: Signal, image and video processing Vol. 18; no. 5; pp. 4713 - 4729
• Main Authors: Prakash, Anju J., Sruthy, S.
• Format: Journal Article
• Language: English
• Published: London Springer London 01.07.2024; Springer Nature B.V
• Subjects: Advanced driver assistance systems; Classification; Computer Imaging; Computer Science; Computer vision; Deep learning; Human error; Image Processing and Computer Vision; Intelligent transportation systems; Machine learning; Multimedia Information Systems; Object recognition; Original Paper; Pattern Recognition and Graphics; Performance evaluation; Roads; Roads & highways; Safety management; Signal,Image and Speech Processing; Traffic control; Traffic information; Traffic management; Traffic safety; Traffic signs; Transportation planning; Vision; Computer vision; Image processing; Driver assistance; Traffic management; Artificial Intelligence; Classification; Road safety; Detection; Intelligent transportation systems
• ISSN: 1863-1703; 1863-1711
• DOI: 10.1007/s11760-024-03108-1

Traffic sign recognition (TSR) systems are essential for strengthening road safety, enhancing traffic management, and promoting efficient driving due to the ever-increasing number of vehicles on the roads and the need for better transportation systems. Modern intelligent transportation systems rely on TSR systems to help with the detection, categorization, and interpretation of traffic signs. These technologies improve driving assistance and road safety by automatically recognizing and comprehending the information contained in traffic signs through the use of computer vision and AI. By automating the recognition process, these systems lessen the load on drivers, limit the possibility of human error, and help to keep the roads safe by reducing infractions and accidents. TSR systems are also necessary for the creation and application of advanced driver assistance systems, laying the foundation for a future that places a premium on safer and more effective transportation. An efficient deep learning (DL)-based model for TSR was presented in this paper. The suggested TSR system consists of two models: a traffic sign detection model that uses RetinaNet and a traffic sign classification model that uses DenseNet-121. The performance evaluation of detection and classification models relies on the utilization of the GTSDB and GTSRB datasets. The performance of the detector was found to surpass that of the existing traffic sign detectors. The traffic sign classifier achieved better classification performance with an accuracy of 98.32%. Using powerful TSR technology could revolutionize transportation and provide a better environment for all drivers.