Shape based image retrieval through fuzzy c means classifier *Ajay B. Kurhe
**Suhas S. Satonkar
***Prakash B. Khanale
*Department of computer science, S.G.B. College, Purna, Dist. Parbhani (M.S.) India. **S.J.College, Gangakhed, Dist. Parbhani (M.S.) India.
**D.S.M. College, Parbhani (M.S.) India.
firstname.lastname@example.org, Suhas.email@example.com, firstname.lastname@example.org
Grouping images into (semantically) meaningful categories using low-level visual features is a challenging and important problem in content-based image retrieval. Using fuzzy C means classifiers, we attempt to capture high-level concepts from low-level image features under the constraint that the test image does belong to one of the classes. Specifically, we consider the hierarchical classification of Coil database images; at the highest level, images are classified as belonged to query image class. We demonstrate that the features used in the classifier are obtained from the directional chain code information of the boundaries of the objects. The bounding box of an object is segmented into four blocks for the spatial relations and the chain code histogram is computed in each of the blocks. Based on the chain code histogram, here we have used 16 dimensional features for recognition. These chain code features are fed to the fuzzy C means classifier for recognition. Best recognition result we obtained is 98%.
We address the problem of two-dimensional (2-D) shape representation and matching for large image databases. The boundary contour of the object must include the boundary part which is entirely inside the outline of the object. For this experiment we used canny function for boundary extraction. In this work chain code is main feature as shown in fig.1 which is basic element of boundary, from which boundaries main parts such as curves, angles, vertical lines, horizontal lines, cross lines are formed. For spatial relations we divide whole boundary into four quadrant. Color and texture contain important information but, for instance, two images with similar color histograms can represent very different things. Therefore the use of shape-describing features is essential in an efficient content-based image retrieval system. Shape image has a value 0 or 1 and we consider the main feature as chain code, which forms the curves, angles and different lines of the shapes. For measuring curves, angles and different lines we used chain code, for spatial relations we divide whole shape into 4 quadrant. We computed histograms of chain code of test image and same histogram computed for database images and computed distances of chain code histogram of test image with chain code histogram of the database images. For the classification we used the fuzzy c means distance, we employed this system on COIL database for 10 objects images in which each object has 72 images. Result of recall is given in below section.
1.2 Fuzzy C-Means Classifier (clustering)
In fuzzy clustering, each point has a degree of belonging to clusters, as in fuzzy logic, rather than belonging completely to just one cluster. Thus, points on the edge of a cluster, may be in the cluster to a lesser degree than points in the center of cluster. An overview and comparison of different fuzzy clustering algorithms is available. In hard clustering, data is divided into distinct clusters, where each data element belongs to exactly one cluster. In fuzzy clustering (also referred to as soft clustering), data elements can belong to more than one cluster, and associated with each element is a set of membership levels. These indicate the strength of the association between that data element and a particular cluster. Fuzzy clustering is a process of assigning these membership levels, and then using them to assign data elements to one or more clusters. Any point x has a set of coefficients giving the degree of being in the kth cluster wk(x). With fuzzy c-means,...
References:  I. Ahmad and W. I. Grosky, (Sept. 2003) “Indexing and retrieval of images by spatial constraint”,
Journal of Visual Communication and Image Representation, 14(3):291–320.
 Imran Ahmad, Muhammad Talal Ibrahim, 2006, “Image Classification and Retrieval using
Correlation”, Proceedings of the 3rd Canadian Conference on Computer and Robot Vision
 Sami Brandt, Jorma Laaksonen and Erkki Oja, (2000), “Statistical Shape Features in Content-
Based Image Retrieval”, 0-7695-0750-6/0$01 0.00 0 IEEE.
 P. B. Khanale and A. B. Kurhe, (2011), “Color Perception of Images using Fuzzy Logic”,
Advances in Computational Sciences and Technology, Volume 4 Number 1pp
 Ahmed, Mohamed N.; Yamany, Sameh M.; Mohamed, Nevin; Farag, Aly A.; Moriarty, Thomas
 Nock, R. and Nielsen, F. (2006) "On Weighting Clustering", IEEE Trans. on Pattern Analysis and
Machine Intelligence, 28 (8), 1–13
 Bezdek, James C. (1981). Pattern Recognition with Fuzzy Objective Function Algorithms. ISBN 0-306-
 A. Pentland, R.W. Picard, and S. Sclaroff, (1994), “Photobook: Tools for Content-Based
Manipulation of Image Databases”, In Storage and Retrieval for Image and Video Databases II,
 A. W. M. Smeulders, M. Worring, S. Santini, A. Gupta, and R. Jain, (Dec. 2000), “Content-based
image retrieval at the end of the early years”, IEEE Transaction on Pattern Analysis and Machine
 J. R. Smith and S.-F. Chang, (1996), “VisualSEEk: A Fully Automated Content-Based Image
Query System”, In Proceedings of ACM Conference on Multimedia, pages 211–218.
 URL. http://www.lems.brown.edu/˜dmc, 2006.
 B. M. Wei-Ying Ma., (1999), “NeTra: A Toolbox for Navigating Large Image Database”,
Multimedia Systems, 7(3):184–198.
Please join StudyMode to read the full document