Moving objects detection in thermal scene videos using unsupervised Bayesian classifier with bootstrap Gaussian expectation maximization algorithm

In this paper, a new algorithm for moving object detection is proposed by using unsupervised Bayesian classifier with bootstrap Gaussian expectation maximization algorithm. It consists of the following steps: the first contains of classify and estimate the motion vectors between successive frames us...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Multimedia tools and applications Jg. 83; H. 2; S. 6335 - 6350
1. Verfasser: Kerfa, Djoudi
Format: Journal Article
Sprache:Englisch
Veröffentlicht: New York Springer US 01.01.2024
Springer Nature B.V
Schlagworte:
Algorithms
Bayesian analysis
Classifiers
Computer Communication Networks
Computer Science
Data Structures and Information Theory
Diamonds
Maximization
Morphology
Moving object recognition
Multimedia
Multimedia Information Systems
Optimization
Pixels
Search algorithms
Special Purpose and Application-Based Systems
Surveillance
Track 6: Computer Vision for Multimedia Applications
Gaussian expectation of maximization
Unsupervised bayesian classifier
Block matching algorithm
Infrared video sequences
Moving object detection
ISSN:1380-7501, 1573-7721
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:In this paper, a new algorithm for moving object detection is proposed by using unsupervised Bayesian classifier with bootstrap Gaussian expectation maximization algorithm. It consists of the following steps: the first contains of classify and estimate the motion vectors between successive frames using the Star diamond search algorithm based on unsupervised Bayesian classifier with Gaussian Expectation of Maximization algorithm, this step serves also to detect the static and dynamic blocks. In the second step, the dynamic blocks are compensated with the white pixels value and the stationary are compensated by black pixels value. In the third step, the morphological opening and closing filters are used for refining the object detected. The proposed approach is trained and evaluated using available infrared (FLIR_ADAS_v2) dataset. The results demonstrate the effectiveness of the proposed method.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1380-7501
1573-7721
DOI:10.1007/s11042-023-15849-1