Influence of image compression on object detection in natural images segmented with mean shift

Authors:
Ana Gaši;Hrvoje Dujmić;Hrvoje Turić;Vladan Papić
Affiliations:
Department of Electronics, University of Split, Croatia;Department of Electronics, University of Split, Croatia;Department of Polytechnics, University of Split, Croatia;Department of Electronics, University of Split, Croatia
Venue:
SIP'10 Proceedings of the 9th WSEAS international conference on Signal processing
Year:
2010

Citing 5
Cited 0

Automatic object extraction from aerial imagery—a survey focusing on buildings

Computer Vision and Image Understanding
Mean Shift: A Robust Approach Toward Feature Space Analysis

IEEE Transactions on Pattern Analysis and Machine Intelligence
State of the art on automatic road extraction for GIS update: a novel classification

Pattern Recognition Letters
Evaluation of Maritime Vision Techniques for Aerial Search of Humans in Maritime Environments

DICTA '08 Proceedings of the 2008 Digital Image Computing: Techniques and Applications
Two-stage segmentation for detection of suspicious objects in aerial and long-range surveillance applications

ICAI'09 Proceedings of the 10th WSEAS international conference on Automation & information

Quantified Score

Hi-index	0.00

Visualization

Abstract

One of important image processing applications is detection of humans for non-urban Search and Rescue. Authors have previously developed a method which, using mean shift segmentation, searches for objects (humans and artificial objects) in natural images. It is based on assumption that such objects have different color comparing to the rest of image. Main problem in using mean shift algorithm is long processing time. In this paper, influence of image compression ratio on mean shift processing time as well as on detection results has been investigated. It is shown that increase in compression ratio results with shorter processing time while detection of objects (Recall and Precision) is not significantly deteriorated.