System for measuring myocardium in cardiac images
First Claim
1. A method of determining epicardial boundary of myocardium of a subject'"'"'s from a cardiac image having pixel intensities at (r,θ
- ) of I(r,θ
) comprising the steps of;
a) identifying a myocardium inner boundary of said image r(θ
) at several discrete angles θ
i from said image by conventional means;
b) determining a goodness function G(I(r,θ
)) being positive if pixel intensity I(r,θ
) at (r,θ
) is deemed to be myocardium, and negative is other tissue, for a plurality of (r,θ
) coordinates;
c) calculating the second derivative and fourth derivatives of radial change due to change in angle, ##EQU7## respectively;
d) calculating a localized energy function H(θ
) from G(I(r,θ
)), ##EQU8## α
, β
, γ
employing calculus of variation;
e) determining δ
r(θ
i) from -ε
H(θ
) where ε
is selected to make the product |-ε
H(θ
)| less than a single pixel width;
f) adding dr(θ
i) to each value of r(θ
i) to determine a new boundary; and
g) repeating steps "c"-"f" for a plurality of iterations to result in an epicardial boundary r(θ
i).
1 Assignment
0 Petitions
Accused Products
Abstract
The present invention determines the epicardial boundary, being a closed curve dividing the myocardium from the tissue and blood surrounding the left ventricle. A mean and standard deviation is determined for pixels of a medical image of the subject'"'"'s myocardial tissue. These are used to define a "goodness function" over the image which is positive for pixels statistically likely to be myocardial tissue, and negative for other pixels. An initial curve for modeling the epicardium in radial coordinates starts with a curve of inner myocardial boundary obtained my conventional imaging techniques. This curve is then iteratively updated to maximize the total "goodness function" of the region encompassed.
-
Citations
3 Claims
-
1. A method of determining epicardial boundary of myocardium of a subject'"'"'s from a cardiac image having pixel intensities at (r,θ
- ) of I(r,θ
) comprising the steps of;a) identifying a myocardium inner boundary of said image r(θ
) at several discrete angles θ
i from said image by conventional means;b) determining a goodness function G(I(r,θ
)) being positive if pixel intensity I(r,θ
) at (r,θ
) is deemed to be myocardium, and negative is other tissue, for a plurality of (r,θ
) coordinates;c) calculating the second derivative and fourth derivatives of radial change due to change in angle, ##EQU7## respectively;
d) calculating a localized energy function H(θ
) from G(I(r,θ
)), ##EQU8## α
, β
, γ
employing calculus of variation;
e) determining δ
r(θ
i) from -ε
H(θ
) where ε
is selected to make the product |-ε
H(θ
)| less than a single pixel width;f) adding dr(θ
i) to each value of r(θ
i) to determine a new boundary; andg) repeating steps "c"-"f" for a plurality of iterations to result in an epicardial boundary r(θ
i). - View Dependent Claims (2, 3)
- ) of I(r,θ
Specification
-
- Resources
-
Current AssigneeGeneral Electric Company
-
Original AssigneeGeneral Electric Company
-
InventorsCurwen, Rupert William Meldrum, Hartley, Richard Ian
-
Primary Examiner(s)JAWORSKI, FRANCIS J
-
Application NumberUS08/751,345Time in Patent Office308 DaysField of Search128/653.1, 128/653.2, 128/659, 128/660.07, 128/661.04, 128/661.01, 128/695, 128/713, 382/128, 382/130-132, 382/266, 382/270-273, 364/413.13, 364/413.25US Class Current600/425CPC Class CodesA61B 2576/023 for the heartG06T 2207/30048 Heart; CardiacG06T 7/0012 Biomedical image inspectionG06T 7/12 Edge-based segmentationG06T 7/149 involving deformable models...G16H 30/40 for processing medical imag...
