Vehicle Accident and Traffic Classification Using Deep Convolutional Neural Networks

Road Traffic accident is a horrifying and terrifying phenomenon which is the main cause of death of more than 1.35 million people and injuries to over 50 million people per year. The global economic loss caused by road traffic accidents annually is estimated at 518 billion US dollars. In this work, we investigate the categorization of traffic and accident images according to their content, features, and semantics using deep learning techniques. The availability of appropriate data representation enhances the performance of deep neural networks in feature extraction using supervised domain knowledge of the data to create valuable pattern understanding. In this work, we have gone through the classification of images using CNN techniques and categorize images into four predefined classes i.e. Accident, Dense Traffic, Fire, and Sparse Traffic. Our deep Convolutional Neural Network (CNN) model learning the mapping of the input images to their labeled classes and show good generalization to the test dataset. Experimental results on real-world datasets have shown that the proposed CNN method is effective for accident and traffic classification, and can perform the task with an accuracy of 94.4% on four target traffic and accident classes.