Open AccessNew Method to Detect Salient Objects in Image Segmentation using Hypergraph Structure
Author(s) -
Eugen Ganea,
Dumitru Dan Burdescu,
Marius Brezovan
Publication year - 2011
Publication title -
advances in electrical and computer engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.254
H-Index - 23
eISSN - 1844-7600
pISSN - 1582-7445
DOI - 10.4316/aece.2011.04018
Subject(s) - salient , artificial intelligence , segmentation , image segmentation , computer vision , computer science , pattern recognition (psychology) , representation (politics) , segmentation based object categorization , image (mathematics) , hypergraph , scale space segmentation , feature detection (computer vision) , image processing , mathematics , combinatorics , politics , political science , law
This paper presents a method for detection of salient objects from images. The proposed algorithms for image segmentation and objects detection use a hexagonal representation of the image pixels and a hypergraph structure to process this hierarchal structure. The main goal of the method is to obtain salient regions, which may be associated with semantic labels. The designed algorithms use color characteristic and syntactic features for image segmentation. The object-oriented model used for storing the results of the segmentation and detection allows directly annotation of regions without a processing of these. The experiments showed that the presented method is robust and accurate comparing with others public methods used for salient objects detection