RT Conference Proceedings
T1 Deep learning-based anomalous object detection system for panoramic cameras managed by a Jetson TX2 board
A1 Benito-Picazo, Jesús
A1 Domínguez-Merino, Enrique
A1 Palomo-Ferrer, Esteban José
A1 Ramos-Jiménez, Gonzalo Pascual
A1 López-Rubio, Ezequiel
K1 Videovigilancia - Congresos
AB Social conflicts appearing in the media are increas ing public awareness about security issues, resulting in a higherdemand of more exhaustive environment monitoring methods.Automatic video surveillance systems are a powerful assistance topublic and private security agents. Since the arrival of deep learn ing, object detection and classification systems have experienceda large improvement in both accuracy and versatility. However,deep learning-based object detection and classification systemsoften require expensive GPU-based hardware to work properly.This paper presents a novel deep learning-based foregroundanomalous object detection system for video streams supplied bypanoramic cameras, specially designed to build power efficientvideo surveillance systems. The system optimises the processof searching for anomalous objects through a new potentialdetection generator managed by three different multivarianthomoscedastic distributions. Experimental results obtained afterits deployment in a Jetson TX2 board attest the good performanceof the system, postulating it as a solvent approach to power savingvideo surveillance systems.
PB IEEE
YR 2021
FD 2021
LK https://hdl.handle.net/10630/30310
UL https://hdl.handle.net/10630/30310
LA eng
NO J. Benito-Picazo, E. Domínguez, E. J. Palomo, G. Ramos-Jiménez and E. López-Rubio, "Deep learning-based anomalous object detection system for panoramic cameras managed by a Jetson TX2 board," 2021 International Joint Conference on Neural Networks (IJCNN), Shenzhen, China, 2021, pp. 1-7, doi: 10.1109/IJCNN52387.2021.9534053
NO This work is partially supported by the Ministry of Economyand Competitiveness of Spain under grants TIN2016-75097-P and PPIT.UMA.B1.2017. It is also partially supported bythe Ministry of Science, Innovation and Universities of Spainunder grant RTI2018-094645-B-I00, project name Automateddetection with low-cost hardware of unusual activities in videosequences. It is also partially supported by the AutonomousGovernment of Andalusia (Spain) under project UMA18-FEDERJA-084, project name Detection of anomalous behavioragents by deep learning in low-cost video surveillance intel ligent systems. All of them include funds from the EuropeanRegional Development Fund (ERDF). The authors thankfullyacknowledge the computer resources, technical expertise andassistance provided by the SCBI (Supercomputing and Bioin formatics) center of the University of Malaga. They also ´Authorized licensed use limited to: Universidad de Malaga. Downloaded on February 06,2024 at 07:21:43 UTC from IEEE Xplore. Restrictions apply. gratefully acknowledge the support of NVIDIA Corporationwith the donation of two Titan X GPUs used for this research.The authors acknowledge the funding from the Universidad deMalaga.
DS RIUMA. Repositorio Institucional de la Universidad de Málaga
RD 19 ene 2026