Co-occurrence of local anisotropic gradient orientations

Patent metadata

Field	Value
Publication number	US-9483822-B2
Application number	US-201514607145-A
Country	US
Kind code	B2
Filing date	Jan 28, 2015
Priority date	Mar 10, 2014
Publication date	Nov 1, 2016
Grant date	Nov 1, 2016

Abstract

Methods, apparatus, and other embodiments associated with distinguishing disease phenotypes using co-occurrence of local anisotropic gradient orientations (CoLIAGe) are described. One example apparatus includes a set of logics that acquires a radiologic image (e.g., MRI image) of a region of tissue demonstrating disease pathology (e.g., cancer), computes a gradient orientation for a pixel in the MRI image, computes a significant orientation for the pixel based on the gradient orientation, constructs a feature vector that captures a discretized entropy distribution for the image based on the significant orientation, and classifies the phenotype of the disease pathology based on the feature vector. Embodiments of example apparatus may generate and display a heatmap of entropy values for the image. Example methods and apparatus may operate substantially in real-time. Example methods and apparatus may operate in two or three dimensions.

First claim

What is claimed is:

1. A non-transitory computer-readable storage medium storing computer-executable instructions that when executed by a computer control the computer to perform a method for distinguishing disease phenotypes using co-occurrence of local anisotropic gradient orientations (CoLIAGe), the method comprising:
accessing a region of interest (ROI) in a volume illustrated in a magnetic resonance image (MRI), the ROI having a set of pixels, and where a pixel in the set of pixels has an intensity;
obtaining an x-axis gradient for a first pixel in the set of pixels based, at least in part, on the intensity of the pixel;
obtaining a y-axis gradient for the first pixel based, at least in part, on the intensity of the pixel;
computing an x-axis gradient vector for a second pixel in an N pixel by N pixel neighborhood centered around the first pixel, N being a number;
computing a y-axis gradient vector for the second pixel in the N pixel by N pixel neighborhood;
constructing a localized gradient vector matrix based, at least in part, on the x-axis gradient vector for the second pixel, and the y-axis gradient vector for the second pixel;
computing a dominant orientation for the first pixel based, at least in part, on the localized gradient vector matrix;
constructing a co-occurrence matrix from the dominant orientation;
computing an entropy for the first pixel based, at least in part, on the co-occurrence matrix;
obtaining a distribution of the entropy;
constructing a feature vector based, at least in part, on the distribution of the entropy; and
controlling a phenotype classifier to classify the ROI based, at least in part, on the feature vector.

2. The non-transitory computer-readable storage medium of claim 1 , where the volume illustrated in the MRI is associated with a Gadolinium-contrast (Gd-C) T1-weighted MRI image of a patient demonstrating brain cancer pathology.

3. The non-transitory computer-readable storage medium of claim 2 , where controlling the phenotype classifier to classify the ROI includes distinguishing radiation necrosis (RN) from recurrent brain tumors (rBT) in the ROI.

4. The non-transitory computer-readable storage medium of claim 1 , where the volume illustrated in the MRI is associated with a dynamic contrast enhanced (DCE)-MRI image of a patient demonstrating breast cancer pathology.

5. The non-transitory computer-readable storage medium of claim 4 , where controlling the phenotype classifier to classify the ROI includes identifying phenotypic imaging signatures of a plurality of molecular sub-types of breast cancer, where the plurality of sub-types includes triple negative (TN), estrogen receptor-positive (ER+), human epidermal growth factor receptor positive (HER2+), and benign fibroadenoma (FA).

6. The non-transitory computer-readable storage medium of claim 1 , where the ROI is defined as =(C,f), where f(c) is an associated intensity at the first pixel c on a three dimensional (3D) grid C.

7. The non-transitory computer-readable storage medium of claim 6 , where obtaining the x-axis gradient for the first pixel and obtaining the y-axis gradient for the first pixel includes computing

∇

f
⁡

(
c
)

=

∂

f
⁡

(
c
)

∂
X

⁢

i
^

+

∂

f
⁡

(
c
)

∂
Y

⁢

j
^

,

where

∂

f
⁡

(
c
)

∂
X

represents the gradient magnitude along the X axis, and
where

∂

f
⁡

(
c
)

∂
Y

represents the gradient magnitude along the Y axis.

8. The non-transitory computer-readable storage medium of claim 1 , where constructing the localized gradient vector matrix comprises computing the localized gradient vector matrix in a two dimensional neighborhood of dimension N 2 ×2.

9. The non-transitory computer-readable storage medium of claim 8 , where computing the x-axis gradient vector for the second pixel in the N pixel by N pixel neighborhood centered around the first pixel includes computing {right arrow over (∂f X )}(c k ), and where computing the y-axis gradient vector for the second pixel includes computing {right arrow over (∂f Y )}(c k ), where kε{1, 2, . . . , N 2 }.

10. The non-transitory computer-readable storage medium of claim 9 , where the localized gradient vector matrix is defined as {right arrow over (F)}=[{right arrow over (∂f X )}(c k ){right arrow over (∂f Y )}(c k )].

11. The non-transitory computer-readable storage medium of claim 10 , where computing the dominant orientation for the first pixel includes calculating

ϕ
⁡

(
c
)

=

tan

-
1

⁢

r
Y
k

r
X
k

,

Assignees

• Univ Case Western Reserve

Inventors

• Madabhushi Anant
• Tiwari Pallavi
• Prasanna Prateek

Classifications

• G06T7/404

  Physics · mapped topic

• G06T7/0012 Primary

  Biomedical image inspection · CPC title

• G06T2207/20081

  Training; Learning · CPC title

• G06T2207/10088

  Magnetic resonance imaging [MRI] · CPC title

• G06T2207/30016

  Brain · CPC title