Please use this identifier to cite or link to this item: http://localhost:8081/xmlui/handle/123456789/15387
Full metadata record
DC FieldValueLanguage
dc.contributor.authorChandel, Rajeev-
dc.date.accessioned2022-04-26T07:00:17Z-
dc.date.available2022-04-26T07:00:17Z-
dc.date.issued2013-06-
dc.identifier.urihttp://localhost:8081/xmlui/handle/123456789/15387-
dc.description.abstractIn this thesis, Fuzzy C-means technique is presented as basis for image segmentation process. The various aspects of working of Fuzzy C-means algorithms are highlighted and the sequential development of these algorithms is given. The advantages of kernel functions and their use in the process of image segmentation are specified. Fuzzy C-Means (FCM) is a prevalent soft-clustering technique. This clustering technique is widely used in the task of image segmentation because of its ease of execution and rapid convergence. By using kernel properties, the Kernel Fuzzy C-Means algorithm attempts to map the data with nonlinear relationships to appropriate higher dimensional spaces. In the new space, the data can be more easily separated or clustered. Kernel combination, or selection, is fundamental for effective kernel clustering. By incorporating multiple kernels and automatically adjusting the kernel weights, Multiple Kernel Fuzzy C-Mean (MKFC) method is more immune to ineffective kernels and their extraneous features. This makes the choice of kernels less crucial and the method more effective for image segmentation. Effective kernels and associated features tend to contribute more to the clustering and, therefore, improve results of image segmentation. The MKFCM algorithm provides us a new platform to blend different types of image information in imagesegmentation problems. In this report, the technique for weight optimization is presented whereby obviating the need for centre calculations for the clusters and improving the performance of Multiple Kernel Fuzzy C-Mean algorithm. Simulations on the segmentation of synthetic, medical image and other images demonstrate the flexibility and advantages of MKFCM based approaches for image segmentation. The values of various parameters involved in the MKFCM algorithm are studied and guidelines for value selection are suggested. The fuzzy c-mean technique for image segmentation is a robust, easy to realize and effective methodology. Apart from these advantages it offers a great benefit by providing a platform for information fusion.en_US
dc.description.sponsorshipINDIAN INSTITUTE OF TECHNOLOGY ROORKEEen_US
dc.language.isoen.en_US
dc.publisherI I T ROORKEEen_US
dc.subjectFuzzy C-meansen_US
dc.subjectEffective Kernel Clusteringen_US
dc.subjectMultiple Kernelen_US
dc.subjectImagesegmentationen_US
dc.titleFUZZY C-MEAN TECHNIQUE FOR IMAGE SEGMENTATIONen_US
dc.typeOtheren_US
dc.accession.numberG22235en_US
Appears in Collections:MASTERS' THESES (Electrical Engg)

Files in This Item:
File Description SizeFormat 
G22235.pdf8.28 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.