Title: HADEM-MACS: a hybrid approach for detection and extraction of objects in movement by multimedia autonomous computer systems

Authors: Elie Tagne Fute; Lionel Landry Sop Deffo; Emmanuel Tonye

Addresses: Department of Computer Engineering, University of Buea, Buea, Cameroon; Department of Mathematics and Computer Science, University of Dschang, Dschang, Cameroon ' Department of Computer Engineering, University of Buea, Buea, Cameroon; Department of Mathematics and Computer Science, University of Dschang, Dschang, Cameroon ' Department of Electrical and Telecommunication Engineering, National Advanced School of Engineering, University of Yaounde I, Yaounde, Cameroon

Abstract: Nowadays, we assist in the multiplication of applications that need to exactly identify objects crossing a field watched by cameras. Multimedia information thus becomes an inescapable medium for the validation of applications such as identification, localisation and object tracking. This gives rise to many processing methods that, after collecting multimedia data (images, videos), continue with a pre-processing in order to reduce noise, finally it finishes with processing in order to extract objects, more precisely the form of object that capture our domain of interest. The first stage consists of detecting objects in movement in the scene. This detection passes through a background modelling. Model based on mixture of Gaussian is commonly used and it is classified in the categories of parametric approaches. However, many other approaches exist today in literature each of them having its advantages and drawbacks. We present in this paper a hybrid approach of movement detection which combines an improved version of mixture of Gaussian with a neural network-based approach.

Keywords: background subtraction; movement detection; mixture of Gaussian; MoG; neural network; multimedia.

DOI: 10.1504/IJCVR.2021.111882

International Journal of Computational Vision and Robotics, 2021 Vol.11 No.1, pp.21 - 40

Accepted: 16 Sep 2019
Published online: 18 Dec 2020 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article