This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: @à _ È c- Ì bAE ¶ W ¿ i7 Š z l å )U :NSC 95-2213-E-134-001 Ï W ‚ Ì : 95 8 ~ 1 n B 96 7 ~ 31 n 3 M A : £ d G Å h- ` > ×ç @àb ç Í Abstract Image segmentation is an important step in any image analysis system. Existing various segmentation methods for MRI have been used to differentiate abnormal and normal tissues. In this paper, we describe a penalized fuzzy clustering algorithm de- signed to segmentation in ophthalmological MRI. Adopting the idea of penalization, we proposed the so-called penalized inter-cluster separation (PICS) by adding a penalty term to the ICS clustering algorithm . Numerical comparisons are made for several fuzzy clusterings according to criteria of accuracy and computational efficiency. The results show that the PICS is impressive. Finally, the PICS algorithm is applied in the segmentation of magnetic resonance image (MRI) of an ophthalmic patient. In these MRI segmentation results, we find that PICS provides useful information as an aid to diagnosis in ophthalmology. Keywords: Fuzzy clustering; Image segmentation; Magnetic resonance image (MRI); Penalized fuzzy c-means; Penalized ICS. 1 1. Introduction MRI segmentation provides important information for detecting a variety of tumors, lesions, and abnormalities in clinical diagnosis. The segmentation can be described as the definition of clusters whose points are associated to similar sets of intensity values in the different images. An efficient analysis of dual echo medical imaging volumes can be derived from a set of different diagnostic volumes carrying complementary information provided by medical imaging technology. The extraction of such volumes from imaging data is said to be segmentation and it is usually performed, in the image space, defining sets of vowels with similar features within a whole dual echo volume. In general, medical images are obtained using different acquisition methods, including x-ray computer tomography (CT), single photon emission tomography (SPET), positron emission tomography (PET), ultra-sound (US), magnetic resonance image (MRI) and magnetic resonance angiographies (MRA), etc. MRI systems are important in medical image analysis. MRI has the multidimensional nature of data provided from either one of two different pulse sequences. MRI segmentation is an important step in any image analysis system. Various segmentation methods for MRI have been used to differentiate abnormal and normal tissues (see , , , , ). Cluster analysis is a method of grouping data with similar characteristics into larger units of analysis. Image segmentation is a way to partition image pixels into similar regions. Thus, a clustering algorithm would naturally be applied in image segmentation....
View Full Document
This note was uploaded on 11/27/2009 for the course IM MA420 taught by Professor Mar,lee during the Spring '09 term at National Taiwan University.
- Spring '09
- The Land