Categorical vehicle classification using Deep Neural Networks

Abstract

Categorical vehicle classification on sequence of image data being a task of great importance and challenging as well for automated traffic surveillance and for the autonomous vehicles building up nowadays. For an autonomous vehicle, it demands for an approach that can precisely detect and classify the vehicle or object around it while moving to avoid any accident. Deep neural networks approach for object detection and classification overpowers other machine learning algorithms that lags in accuracy and computational complexity. In this paper we have classified different categories of vehicles as HMV, LMV and Two-wheelers using deep neural networks. The implementation utilizes nine-layer network to faithfully detect and classify the vehicle category. The results showcase that we can rely on the deep neural networks approach for the vehicle category detection and classification for the urban traffic surveillance and for the autonomous vehicles.