Colour Based Traffic Sign Detection and Recognition using Support Vector Machine (SVM), Convolutional Neural Networks (CNN)

Rajesh V, Dr. Rajashekar J S

Todays world uses artificial intelligence to automate nearly every field, simplifying everything. This decreases risk while simultaneously increasing accuracy, dependability, and minimizing physical contact. While operating a car, it is crucial to adhere to traffic laws and regulations, but occasionally drivers fail to see the signs that are posted along the sides of the road. This can be fatal for both drivers and pedestrians. Hence the system has to be integrated with ADAS system, which will automatically send the driver a text or voice message to warn them of an approaching traffic sign. In this paper we proposed the model which will detect the traffic sign in diverse background using color information and SVM. Detected Traffic sign is recognized and classified using convolutional neural network. We used Lenet-5 CNN architecture and was found more efficient about traditional CNN model. Thus, making it desirable to apply in real-time computer vision tasks. Proposed design has high detection rate and it is having less complexity

Title:	Colour Based Traffic Sign Detection and Recognition using Support Vector Machine (SVM), Convolutional Neural Networks (CNN)
Author Name:	Rajesh V, Dr. Rajashekar J S
Abstract:	Todays world uses artificial intelligence to automate nearly every field, simplifying everything. This decreases risk while simultaneously increasing accuracy, dependability, and minimizing physical contact. While operating a car, it is crucial to adhere to traffic laws and regulations, but occasionally drivers fail to see the signs that are posted along the sides of the road. This can be fatal for both drivers and pedestrians. Hence the system has to be integrated with ADAS system, which will automatically send the driver a text or voice message to warn them of an approaching traffic sign. In this paper we proposed the model which will detect the traffic sign in diverse background using color information and SVM. Detected Traffic sign is recognized and classified using convolutional neural network. We used Lenet-5 CNN architecture and was found more efficient about traditional CNN model. Thus, making it desirable to apply in real-time computer vision tasks. Proposed design has high detection rate and it is having less complexity
Cite this article:	Rajesh V, Dr. Rajashekar J S , " Colour Based Traffic Sign Detection and Recognition using Support Vector Machine (SVM), Convolutional Neural Networks (CNN) " , International Journal of Emerging Trends & Technology in Computer Science (IJETTCS), Volume 11, Issue 5, September - October 2022 , pp. 013-016 , ISSN 2278-6856.
Full Text [PDF] Home