To identify and recognize the object for traffic analysis system using deep learning

Abstract

The object identification has been most essential field in development of machine vision which should be more efficient and accurate. Machine Learning & Artificial Intelligence, both are on their peak in today’s technology world. Playing with these can leads towards development. The field has actually replaced human efforts. With the approach of profound learning systems (i.e. deep learning techniques), the precision for object identification has expanded radically. This project aims to implement Object Identification for Traffic Analysis System in real time using Deep Learning Algorithms with high accuracy. The differentiation among objects such as humans, Traffic signs, etc. are identified. The dataset is so designed with specific objects which will be recognized by the camera and result will be shown within seconds. The project purely based on deep learning approaches which also includes YOLO object detection & Covolutionary Neural Network (CNN). The resulting system is fast and accurate, therefore can be implemented for smart automation across global stage.