Title: Multiple object clustering using FCM and K-means algorithms

Authors: Sanjivani Shantaiya; Kesari Verma; Kamal K. Mehta

Addresses: Department of Computer Science and Engineering, Disha Institute of Management and Technology, Raipur, India ' Department of Masters of Computer Application, National Institute of Technology, Raipur, India ' Department of Computer Science and Engineering, Institute of Technology, Nirma University, Ahmedabad, India

Abstract: Automatic classification and recognition of images and video is one of the challenging tasks of digital image processing. This paper presents performance analysis of K-means and fuzzy K-means clustering algorithms along with experimental study. The objectives of this paper are automatic extraction of features from images of vehicles and pedestrians and classify them. A good set of features that capture the most important properties of an object are used to identify the objects uniquely. The objects are classified into three different clusters such as pedestrians, light vehicles and heavy vehicles. The experimental studies were performed in MATLAB for K-means and c-means clustering algorithms. K-means clustering proven to be more effective than fuzzy c-means clustering algorithm.

Keywords: object detection; classification; fuzzy c-means clustering; FCM; K-means clustering; multiple objects; digital images; image processing; feature extraction; pedestrians; light vehicles; heavy vehicles.

DOI: 10.1504/IJCVR.2016.079395

International Journal of Computational Vision and Robotics, 2016 Vol.6 No.4, pp.331 - 343

Received: 15 Aug 2014
Accepted: 07 Oct 2014
Published online: 28 Sep 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article