Vehicle Activity Recognition Using DCNN

AlZoubi, Alaa and Nam, David (2020) Vehicle Activity Recognition Using DCNN. In: Computer Vision, Imaging and Computer Graphics Theory and Applications. VISIGRAPP 2019. Communications in Computer and Information Science (1182). Springer, Cham, pp. 566-588. ISBN 978-3-030-41589-1

[img]
Preview
Text
Vehicle Activity Recognition Using DCNN.pdf

Download (2MB) | Preview
Official URL: https://link.springer.com/chapter/10.1007/978-3-03...

Abstract

Abstract. This paper presents a novel Deep Convolutional Neural Net work (DCNN) method for vehicle activity classification. We extend our previous approach to be able to classify a larger number of vehicle trajec tories in a single network.We also highlight the fl exibility of our approach in integrating further scenarios to our classifier. Firstly, a spatiotempo ral calculus method is used to encode the relative movement between vehicles as a trajectory of QTC states. We then map the encoded trajectory to a 2D matrix using the one-hot vector mapping, this preserves the important positional data and order for each QTC state. To do this we associate the QTC sequences with pixels to form a 2D image tex ture. Afterwards, we adapted trained CNN architecture into our vehicles activity recognition task. Two separate types of driving data sets are used to evaluate our method. We demonstrate that the proposed method out-performs existing techniques. Along with the proposed approach we created a new dataset of vehicles interactions. Although the focus of this paper is on the automated analysis of vehicle interactions, the proposed technique is general and can be applied for pairwise analysis for moving objects.

Item Type: Book Section
Uncontrolled Keywords: Vehicle Activity Classifcation; Spatiotemporal Calculus; Trajectory Texture; Transfer Learning; Deep Convolutional Neural Networks
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions: School of Computing
Depositing User: Alaa AlZoubi
Date Deposited: 11 Mar 2020 15:59
Last Modified: 20 Feb 2022 01:15
URI: http://bear.buckingham.ac.uk/id/eprint/459

Actions (login required)

View Item View Item