Title page for ETD etd-1106103-120852


Type of Document Master's Thesis
Author Balasubramanian, Madhusudhanan
Author's Email Address mbalas1@lsu.edu
URN etd-1106103-120852
Title Computer Assisted Eye Fungal Infection Diagnosis
Degree Master of Science in Systems Science (M.S.S.S.)
Department Computer Science
Advisory Committee
Advisor Name Title
Louise A. Perkins Committee Chair
Donald H. Kraft Committee Member
S. Sitharama Iyengar Committee Member
Keywords
  • adaptive mixtures
  • fisher linear discriminant
  • fractal-based features
  • fungal keratitis
  • fungal infection
Date of Defense 2002-11-14
Availability unrestricted
Abstract
In this thesis, an attempt has been made to assist the diagnosis of Fungal Keratitis, a fungal infection that occurs in the corneal layers of the eye, by identifying the region of infection in the corneal images using fractal-based features. Three features related to the fractal dimension of the surface of the image, when represented in a 3D using the pixel intensity measure, are used to identify these regions in the image. To reduce the computation complexity, Fisher linear discriminant (FLD) is used to reduce the 3D raw feature to 1D feature, while preserving feature values. Using the adaptive mixtures (AM) method, the probability density distribution of the two class fractal features, is estimated. A training corneal image has been used to build the two-class probability density distribution. In this work, we use Bayesian classifier, a standard statistical pattern classification technique, to classify the pixels in corneal images, using the two-class probability density distribution. The classifier outputs an image mask, highlighting the fungal infected region in the corneal image. The whole system is implemented in MATLAB.
Files
  Filename       Size       Approximate Download Time (Hours:Minutes:Seconds) 
 
 28.8 Modem   56K Modem   ISDN (64 Kb)   ISDN (128 Kb)   Higher-speed Access 
  Balasubramanian_thesis.pdf 1.93 Mb 00:08:55 00:04:35 00:04:01 00:02:00 00:00:10

Browse All Available ETDs by ( Author | Department )

If you have questions or technical problems, please Contact LSU-ETD Support.