Automatic Herniation Detection A collaborative project with Doug Dean Erin Hannen Purpose Development of an algorithm for S egmentation of individual spinal disks Determination of specific quantitative properties of each disc ID: 267348
Download Presentation The PPT/PDF document "Quantitative Comparison of Conventional ..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Quantitative Comparison of Conventional and Oblique MRI for Detection of Herniated Discs
Automatic Herniation Detection
A collaborative project with
Doug Dean
Erin
HannenSlide2
Purpose
Development of an algorithm for:
S
egmentation
of individual spinal
disks
Determination of specific, quantitative properties of each disc
Use properties to determine if a disc is herniated or normalSlide3
Approach
Modify methods from “Desiccation diagnosis in lumbar discs from clinical MRI with a probabilistic model”
Intensity:
Obtain histogram. Herniated discs typically have lower intensity profile due to spreading of the nucleus
pulposus
over a larger area. Individual intensity values and the average intensity value are obtainedSlide4
2. Probabilistic Model:
Z[n] = normalization factor
β
1
,
β
2
= tuning parametersUA = appearance parameter, Determined by intensity values, averaged intensity, and a defined pixel neighborhoodUS = shape parameters Determined based on location coordinates from points that define shape of the disk
ApproachSlide5
Current Progress
Lumbar discs
segmented
Semi-automated
Edge detection, MATLAB image processing tools
Disc location defined
Centroid
of segmentation boundary calculated Overlay of segmentation boundary onto original imageAverage intensity over entire segmented diskSlide6
I = 80.400
I = 86.4614
I = 84.6678
I = 70.1894 Slide7
Project timeline
April 12: First project presentation
April 13-27: Continue reading literature articles comparing methods for disc quantification. Begin writing MATLAB code for herniation detection using data from class labs or phantom images.
April 28: Mid project presentation
April 28-May 15:
Refine segmentation methods, where needed
Finish
developing herniation detection codeEnsure successful implementation using acquired MRI data May 16 & 17: Final project presentation