Recognition of Color Objects Using Hybrids Descriptors

Recognition of Color Objects Using Hybrids Descriptors

Driss Naji (Informations Processing and Telecommunication Teams, Faculty of Science and Technology, Sultan Moulay Slimane University, Beni-Mellal, Morocco), M. Fakir (Processing and Telecommunication Teams, Faculty of Science and Technology, Sultan Moulay Slimane University, Beni-Mellal, Morocco), B. Bouikhalene (Processing and Telecommunication Teams, Faculty of Science and Technology, Sultan Moulay Slimane University, Beni-Mellal, Morocco) and M. Boutaounte (Faculty of Science and Technology, Sultan Moulay Slimane University, Beni-Mellal, Morocco)
Copyright: © 2013 |Pages: 9
DOI: 10.4018/ijcvip.2013100105
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

In this paper, the authors came up with a different approach based on the combination of the different descriptors. For object recognition, regardless of orientation, size and position, feature vectors are computed with the help of Zernike moments and Centrist descriptors. For a large data base the fact of using the classic descriptors has never been a satisfying method for perfect recognition rates. The authors deduced that the combination of descriptors can have good recognition rates, accordingthe result of a comparative study of the different descriptors and the different combinations (Zernike + Centrist, Zernike + ACP, Centrist + ACP). The Zernike moment with Centrist descriptors ended up being the best hybrid description. For the recognition process, the authors opted for support vector machine (SVM) and Neural Networks (NN). The authors illustrate the proposed method on 3D objects using representations of two-dimensional images that are taken from different angles of view are the main features leading the authors to their objective.
Article Preview

1. Introduction

The major task in object recognition is to identify if any, of an object from the set of known objects appear in the given image or image sequence. Color and invariant object recognition is a critical problem in image processing. The feature extraction is a relevant problem because it seems that there are no attributes that can model a basis on all points of view.

In order to improve discrimination and classification of objects several approaches are proposed in the literature, often based on decision theory, feature selection, optimization, learning like (Kim Y K, 1995) using selection of K neighbors by considering the distance difference and the membership grade each neighbor, and (C. Maaoui, 2005) using combination of the two information: the color and shape for the detection and recognition of objects in an image, as well(J. P. Gauthier, 1991) a family of invariants, called Motion Descriptors, which are invariants in translation, rotation, scale and reflexions, (Liefeng Bo, 2011) proposed a hierarchical kernel descriptors that apply kernel descriptors recursively to form image-level features and thus provide a conceptually simple and consistent way to generate image-level features from pixel attributes.

Our goal here is to demonstrate empirically the ability of these descriptors to be successfully used in object recognition. We propose an approach based on the combination of two descriptors. We studied Centrist descriptor introduced by Wu and Rehg Christensen (J. Wu et al., 2009) combined with Zernike Moments introduced by F. Zernike in 1934 (F. Zernike et al 1934). Experiments are conducted using COIL-100 databases. The flow-chart of recognition system adapted is illustrated in (Figure 1).

Figure 1.

Flow-chart of recognition system

The organization of this paper is as follows:

  • Section 2 deals the SVM and NN as method used for classification.

  • Section 3 describes the features extraction method.

  • Section 4 deals with the recognition phase and experiment results.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 7: 4 Issues (2017): 3 Released, 1 Forthcoming
Volume 6: 2 Issues (2016)
Volume 5: 2 Issues (2015)
Volume 4: 2 Issues (2014)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing