|
 About Us
Mission & Vision
Faculty
Staff
Students
Computing Facilities
News
Seminars
Contact Us
Research
Laboratories
CCT
Academics
Admission
Graduate
Undergraduate
Courses
Rm Reservation
|
|
Final Examination for Ph.D. for Hua Cao |
Presentation Title: "A Novel Automated Approach of Multi-modality
Feature-based Registration and Area-based Fusion of Retinal Images"
Graduate Committee:
- Dr. S.S. Iyengar (Major Professor)
- Dr. Nathan Brener
- Dr. Rajgopal Kannan
- Dr. James Catallo
- Dr. Ye-Sho Chen
Date: Nov 9th, 2007
Time: 11:30 AM
Location: Room 256, Coates Hall
Abstract:
Biomedical image registration and fusion are usually scene dependent, and
require intensive computational effort. We have successfully developed a
novel approach of feature-based registration and area-based fusion of
retinal images. Our algorithm, which is reliable, robust, and
time-efficient, has an automatic adaptation from frame to frame with few
tunable threshold parameters. The reference and the to-be-registered images
are from two different modalities, i.e. angiogram grayscale images and
fundus color images. The relative study of retinal images enhances the
information on the fundus image by superimposing information contained in
the Angiogram image. Our registration part is based on retinal vasculature*
extraction using Canny Edge Detector, and identifying control points at
the global direction change pixels using *adaptive exploratory algorithm*.
Shape similarity criteria are employed to match the control points. Our
fusion part is Mutual Pixel Count (MPC) based optimization procedure, which
adjusts the initially selected control points at the sub-pixel level. We
achieved a global maxima equivalent result by calculating MPC local maxima
with an efficient computation cost. The iteration stops either when MPC
reaches the maximum value, or when the maximum allowable loop count is
reached. To our knowledge, it is the first time that the MPC concept has
been introduced into biomedical image fusion area as the measurement
criteria for fusion accuracy. The fusion image is generated based on the
current control point coordinates when the iteration stops. The comparative
study of our automatic registration and fusion scheme against other existing
approaches has shown the advantage of our approach in terms of novelty,
efficiency, and accuracy.
We made two new contributions:
1. Our automatic registration algorithm identifies control points using
adaptive exploratory algorithm. By locating control points at the global
direction change pixel, local direction changes are efficiently avoided.
2. Our automatic optimization fusion algorithm has introduced
Mutual-Pixel-Count (MPC) concept into biomedical image fusion area. We
achieved a global maxima equivalent fusion result by calculating local
maxima of MPC with an efficient computation cost.
All are welcome!
|
|
