Real-Time Object Tracking by CUDA-accelerated Neural Network

Mikhail S. Tarkov^1, and Sergey V. Dubynin²

¹A.V. Rzhanov’s Institute of Semiconductor Physics SB RAS, Novosibirsk, Russia

²Novosibirsk State University, Novosibirsk, Russia

Cite this paper:
Mikhail S. Tarkov and Sergey V. Dubynin. Real-Time Object Tracking by CUDA-accelerated Neural Network. Journal of Computer Sciences and Applications. 2013; 1(1):1-4. doi: 10.12691/jcsa-1-1-1

Abstract

An algorithm is proposed for tracking objects in real time. The algorithm is based on neural network implemented on GPU. Investigation and parameter optimization of the algorithm are realized. Tracking process has accelerated by 10 times and the training process has accelerated by 2 times versus to the sequential algorithm version. The maximum resolution of the frame for real-time tracking and the optimum frame sampling from a movie are calculated.

Keywords:
object tracking neural network parallel computing CUDA

This work is licensed under a Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Figures

References:

[1]	Ahmed, J., Jafri, M. N., Ahmad, J., and Khan, M. I., “Design and Implementation of a Neural Network for Real-Time Object Tracking,” Engineering and Technology, 6. 209-212. 2005.
[2]	Haykin, S. Neural Networks. A Comprehensive Foundation, Prentice Hall Inc., 1999.
[3]	David, B. Kirk, and Wen-mei, W. Hwu, Programming Massively Parallel Processors. A Hands-on Approach, NVIDIA Corporation, 2010.
[4]	Mobileye Advanced Driver Assistant System. URL: http://www.mobileye.com
[5]	Automatic vehicle driving. URL: http://www.argo.ce.unipr.it/ARGO/english/
[6]	Maya – 3D Animation – Autodesk. URL: http://usa.autodesk.com/maya/
[7]	Bradski, G., Kaehler, A., Learning OpenCV, O’Reilly Media, Inc., USA, 2008.