"TRAFFIC SIGNS DETECTION AND RECOGNITION USING DEEP LEARNING WITH 96 PERCENT ACCURACY"

Authors

Kirit Bhalsod
Gujarat Technological University, Chandkheda

Dr. Dipesh Kamdar (Assistant Professor)
V.V.P. Engineering College, Rajkot

Dr. Navneet Ghedia (Principal)
Sanjaybhai Rajguru College of Engineering, Rajkot

Palakkumar Bhatt (Lecturer)
A.V. Parekh Technical Institute, Rajkot

Abstract

A framework is proposed to detect and recognize symbols and text-based traffic guide panels captured in highway environments. This framework could help deliver the text and symbol information from guide panels to human drivers as head-up display information, in this proposal, detection and classification of traffic signs on captured frames are studied. This set is separated into 10 groups: Cluster, Deformed, Mixed, Different Shapes, Illumination, Occlusion, Rotation, Shadows, Sizes, Translation. The implementation should work as fast as the video stream.