Foreground Extraction of Surveillance Video Under Complex Background

In order to overcome the shortcoming of the origin ViBe algorithm, such as ghosting effects caused by moving objects in first frame, we present a novel foreground object extraction method based on an improved ViBe algorithm. In particular, we improve the initialization of ViBe to pick up more backgr...

Full description

Saved in:
Bibliographic Details
Published in:2018 IEEE 8th Annual International Conference on CYBER Technology in Automation, Control, and Intelligent Systems (CYBER) pp. 445 - 450
Main Authors: Yu, Yue, Pan, Hong, Li, Xinde, Lin, Jiaxi
Format: Conference Proceeding
Language:English
Published: IEEE 01.07.2018
Subjects:
camera jitter
Cameras
foreground extraction
Gray-scale
grayscale projection
Heuristic algorithms
Information filters
Intelligent systems
Jitter
Projection algorithms
Robustness
Surveillance
surveillance video
ViBe
Videos
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In order to overcome the shortcoming of the origin ViBe algorithm, such as ghosting effects caused by moving objects in first frame, we present a novel foreground object extraction method based on an improved ViBe algorithm. In particular, we improve the initialization of ViBe to pick up more background information after the foreground object moving around the initialized samples. Then we use a median filter to reduce false detections. Experimental results show that our method can eliminate the ghosting effects effectively. For degraded surveillance video clips with camera jitter, we first apply the grayscale projection algorithm to stabilize the image and then use the proposed method to extract foreground object. Experimental results on some real-world surveillance video clips demonstrate the robustness of the proposed foreground extraction algorithm under different circumstances.
DOI:10.1109/CYBER.2018.8688357