DRIVERS ASSISTANCE FOR TRAFFIC SIGN RECOGNITION USING DEEP LEARNING

Road safety is a paramount concern worldwide, with traffic signs playing a crucial role in guiding drivers and pedestrians and preventing accidents.However, individuals with visual impairments or those distracted while driving may overlook or misinterpret traffic signs, leading to potential hazards.To address this challenge, this project proposes the development of a Voice Alarm System for Traffic Signs is designed to detect and classify traffic signs dynamically and alert drivers via audio cues.The core architecture leverages a hybrid deep learning approach combining Convolutional Neural Networks (CNN) and MobileNet.CNNs are utilized to capture detailed spatial features and robust patterns from the traffic imagery, while the MobileNet framework ensures lightweight, high-speed processing ideal for resource-constrained edge devices and real-time vehicular deployment.The framework evaluates the model's robustness by utilizing the widely recognized German Traffic Sign Recognition Benchmark (GTSRB) dataset, the model achieves a classification accuracy of over 99% with an ultra-low inference footprint, establishing its readiness for embedded real-time automotive execution.Ultimately, this research validates that depth wise model down-scaling provides a highly dependable and computationally economical solution for driver-assistance for traffic sign recognition.