Classification of Traffic Sign for Smart Vehicle using Ensemble Learning and Deep Learning

Traffic signs are crucial for directing traffic, enforcing safe driving practices, and lowering the number of collisions, injuries, and fatalities. Automated traffic sign recognition and identification is a need for all Intelligent Transportation Systems (ITS). It cannot be overstated how important automatic traffic sign detection and identification is in the era of self-driving cars [1]. This will help the driver's eyesight and maneuvering abilities, which can't be relied upon to always be accurate for all weather-related issues. Road safety is being compromised by the human eye's limited ability to read a traffic sign. This study presents a deep-learning-based autonomous traffic sign cognizance technique for India. AlexNet, VGG-16, ResNet-50, and GoogleNet are the four designs of the Convolutional Neural Network (CNN), which is the foundation for autonomous traffic sign detection and identification. The proposed idea was evaluated using a novel dataset made up of 13971 photos divided into 59 categories. As a pre-trained transfer learning model ResNet-50 has greater accuracy and operational speed to produce a best fit model, we will be incorporating these images to it[2]