BREAST TISSUE DENSITY MEASURE

• US 20110013819A1
• Filed: 05/26/2010
• Published: 01/20/2011
• Est. Priority Date: 02/10/2006
• Status: Active Grant

• Current Assignee: BioClinica, Inc.
• Original Assignee: Nordic Bioscience Imaging A/S
• Inventors: Karemore, Gopal R., Brandt, Sami S., Loog, Marco, Raundahl, Jakob, Nielsen, Mads
• Granted Patent: US 8,285,019 B2
• Time in Patent Office: Days
• US Class Current: 382/132
• CPC Class Codes: G06F 18/23213 with fixed number of clusters, e.g. K-means clustering
• CPC Class Codes: G06T 2207/10116 X-ray image
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 2207/30068 Mammography; Breast
• CPC Class Codes: G06T 7/0016 involving temporal comparison
• CPC Class Codes: G06V 10/52 Scale-space analysis, e.g. wavelet analysis multi-scale boundary representations G06V10/42
• CPC Class Codes: G06V 10/763 Non-hierarchical techniques, e.g. based on statistics of modelling distributions
• CPC Class Codes: G06V 2201/03 Recognition of patterns in medical or anatomical images

Abstract

A method of processing a mammogram image to derive a value for a parameter useful in detecting differences in breast tissue in subsequent images of the same breast or relative to a control group of such images, said derived parameter being an aggregate probability score reflecting the probability of the image being a member of a predefined class of mammogram images, comprises computing for each of a multitude of pixels within a large region of interest within the image a pixel probability score assigned by a trained statistical classifier according to the probability of said pixel belonging to an image belonging to said class, said pixel probability being calculated on the basis of a selected plurality of features of said pixels, and computing said parameter by aggregating the pixel probability scores over said region of interest. Saud features may include the 3-jet of said pixels.

Claims

1. A method for ascribing coordinates to positions of points within an image performed in a suitably programmed computer and comprising for each point for which a set of coordinates is desired computing a first coordinate value s and a second coordinate value φ

defining the position of said point in a coordinate system (s, φ

) wherein said coordinate system comprises a set of curves diverging from an apex point lying on a reference line which extends in a reference direction, each curve in said set of curves extending to intersect a base line extending from a first point to a second point, said first coordinate s representing a measure of a geodesic distance from said apex point along one said curve which contains said point, and said second coordinate φ

representing an angle made between said curve and said reference line at said apex point.

2. A method as claimed in claim 1, wherein within a bounded domain sε

S, φ

ε

Φ

said curves are non-crossing.

3. A method as claimed in claim 1, wherein said first coordinate s is equal to or proportional to a ratio a:

b where

a=the geodesic distance of the point from the apex

b=the geodesic distance from the apex to the intersection of said one curve with the base line.

4. A method as claimed in claim 1, wherein said set of curves has 3 degrees of freedom.

5. A method as claimed in claim 4, wherein said curves in said set of curves are parabolas.

6. A method as claimed in claim 1, wherein said image is of a human breast including a nipple and a pectoral muscle, and said apex point is set at said nipple and said base line approximately follows the pectoral muscle or its boundary.

7. A method as claimed in claim 6, wherein said images are mammogram images.

8. A method of comparing information contained within a first image of an object with information contained within at least one related other image,

wherein said first and other images are related by being images from approximately the same perspective of the same object or being images of a second instance of the same type of object,

comprising ascribing coordinates to positions within each image by a method as claimed in claim 1 and comparing at least one characteristic of points in the respective images which have matching s and φ

coordinates.

9. A method as claimed in claim 8, wherein each image is a mammogram image.

10. A method as claimed in claim 8, comprising calculating Gaussian derivatives of a pre-selected order of image intensities at said points.

11. A method as claimed in claim 8, further comprising classifying a said first image as belonging to a first class or a second class by applying a statistical classifier to said point image intensities.

12. A method as claimed in claim 11, wherein said images are mammograms and said classification is between belonging to a group of images of healthy breasts and belonging to a group of images of breasts characterised by cancer.

14. A method as claimed in claim 12, wherein said trained classifier is trained by unsupervised learning.

15. A method as claimed in claim 12, wherein said trained classifier is trained by supervised learning.

16. A method as claimed in claim 14, wherein said classifier is trained by supervised learning based on a set of images associated with a higher breast density and a set of images associated with a lower breast density.

17. A method as claimed claim 12, wherein said pixels are scored as belonging to one of said classes.

18. A method as claimed in claim 12, wherein said pixels are scored according to their probability of belonging to at least one of said classes.

19. A method as claimed in claim 12, wherein said quotient value is determined for each said pixel at each of a plurality of scales in coordinate directions defined by the s and φ

coordinates of said pixel.

20. A method as claimed in claim 18, wherein said quotient values are determined as the normalised difference between eigenvalues of a Hessian matrix based on Gaussian derivatives in s and φ

coordinate directions at a predetermined scale of pixels of the image, which Gaussian derivatives relate the intensity of each pixel to the intensities of the neighbours of said pixel in s and φ

coordinate directions.

21. A method as claimed in claim 12, comprising

plotting said quotient values on a pre-determined model of a cluster map, said cluster map having at least two pre-plotted points;

clustering the plotted values around said at least two pre-plotted points to result in two discrete clusters; and

deriving said parameter that changes alongside changes in breast density by computing the difference in area between said two clusters.

22. A method as claimed in claim 20, wherein the pre-determined model is defined in 3-dimensional space in which the three dimensions respectively represent the quotient value when calculated at a respective one of three different scales.

23. A method as claimed in claim 20, wherein the step of clustering further comprises:

calculating which of said at least two pre-plotted points each plotted value is closest to; and

assigning each value accordingly to said closest pre-plotted point to result in said two discrete clusters of plotted points.

24. A method as claimed in claim 22, comprising a preliminary step of preparing said pre-determined model of the cluster map incorporating said at least two pre-plotted points.

25. A method as claimed in claim 22, wherein the step of preparing the model of the cluster map comprises:

a) computing a quotient value for pixels in a test image of a breast based on the results of Hessian matrices compiled for said pixels in the image;

b) plotting each quotient value in 3-dimensional space;

c) randomly selecting at least two points amongst the quotient values as starting points;

d) calculating the distance between each quotient value and each starting point and assigning each quotient value to the starting point to which it is nearest;

e) calculating a mean of each group of points and redefining the starting point as the calculated mean point;

iteratively repeating steps d) and e)until the calculated mean point is equal to the starting point, whereby the final calculated mean points are arranged to be the pre-plotted points of the model of the cluster map.

26. A method as claimed in claim 22, wherein the number of pre-plotted points determines the number of resulting clusters.

27. A method as claimed in claim 22, wherein the pre-determined model of the cluster map has four pre-plotted points.

28. A method as claimed in claim 26, wherein four points are randomly selected as starting points to result in said four pre-plotted points for the model of the cluster map.

29. A method as claimed in claim 19, wherein the Hessian matrices are derived from Gaussian derivatives of the pixels in the image.

30. A method as claimed in claim 28, further comprising deriving Gaussian derivates at three different scales of the image to result in three different quotient values for each of said pixels, wherein the three quotient values correspond to the three dimensions of the pre-determined model.

31. A method as claimed in claim 12, wherein said quotient values define characteristics representative of the shape of objects present in the image and a quotient value of relatively large magnitude represents a substantially elongate object located in the image.

13. A method of processing a mammogram image which includes a nipple and a chest wall to derive a value for a parameter useful in detecting differences in breast tissue in subsequent images of the same breast or relative to a control group of such images, said derived parameter being a parameter that changes alongside changes in breast density, the method comprising the steps of processing an image of at least part of a breast by:

calculating in a suitably programmed computer for pixels of the image respective coordinates by a method comprising for each point for which a set of coordinates is desired computing a first coordinate value s and a second coordinate value φ

defining the position of said point in a coordinate system (s, φ

) wherein said coordinate system comprises a set of curves diverging from an apex point set at said nipple and lying on a reference line which extends in a reference direction, each curve in said set of curves extending to intersect a base line set generally to follow said chest wall and extending from a first point to a second point,

said first coordinate s representing a measure of a geodesic distance from said apex point along one said curve which contains said point, and said second coordinate φ

representing an angle made between said curve and said reference line at said apex point

computing for pixels of the image a quotient value representative of the aspect ratio of tissue structures depicted in the image;

using a trained classifier to classify said pixels according to their respective said quotient values and assigning a score to the respective pixels representing their classification with respect to at least two classes;

deriving said parameter that changes alongside changes in breast density based on the aggregate pixel membership scores of said classes.