System and method for fetal biometric measurements from ultrasound data and fusion of same for estimation of fetal gestational age

• US 7,783,095 B2
• Filed: 08/09/2006
• Issued: 08/24/2010
• Est. Priority Date: 08/11/2005
• Status: Active Grant

• Current Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Original Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Inventors: Carneiro, Gustavo, Georgescu, Bogdan, Favaro, Paolo, Comaniciu, Dorin, Good, Sara
• Primary Examiner(s): Bella; Matthew C
• Assistant Examiner(s): Rosario; Dennis
• Application Number:
  US11/463,341
• Publication Number:
  US 20070081705A1
• Time in Patent Office: 1,476 Days
• Field of Search: 382/128
• US Class Current: 382/128
• CPC Class Codes: A61B 8/0866 involving foetal diagnosis; pre-natal or peri-natal diagnosis of the baby
• CPC Class Codes: A61B 8/483 involving the acquisition of a 3D volume of data
• CPC Class Codes: A61B 8/523 for generating planar views from image data in a user selectable plane not corresponding to the acquisition plane
• CPC Class Codes: G06F 18/2321 using statistics or function optimisation, e.g. modelling of probability density functions
• CPC Class Codes: G06F 18/251 of input or preprocessed data
• CPC Class Codes: G06F 18/254 of classification results, e.g. of results related to same input data
• CPC Class Codes: G06T 2207/30004 Biomedical image processing
• CPC Class Codes: G06T 7/62 of area, perimeter, diameter or volume
• CPC Class Codes: G06V 10/763 Non-hierarchical techniques, e.g. based on statistics of modelling distributions
• CPC Class Codes: G06V 10/803 of input or preprocessed data

Abstract

A method for segmenting and measuring anatomical structures in fetal ultrasound images includes the steps of providing a digitized ultrasound image of a fetus comprising a plurality of intensities corresponding to a domain of points on a 3-dimensional grid, providing a plurality of classifiers trained to detect anatomical structures in said image of said fetus, and segmenting and measuring an anatomical structure using said image classifiers by applying said elliptical contour classifiers to said fetal ultrasound image, wherein a plurality of 2-dimensional contours characterizing said anatomical structure are detected. The anatomical structure measurement can be combined with measurement of another anatomical structure to estimate gestational age of the fetus.

Claims

1. A method for segmenting and measuring anatomical structures in fetal ultrasound images comprising the steps of:

providing a digitized 3-dimensional ultrasound image of a fetus comprising a plurality of intensities corresponding to a domain of points on a 3-dimensional grid;

providing a plurality of classifiers trained to detect anatomical structures in said image of said fetus;

segmenting and measuring an anatomical structure using said image classifiers by applying elliptical contour classifiers to said fetal ultrasound image, wherein a plurality of 2-dimensional contours characterizing said anatomical structure are detected; and

combining said anatomical structure measurement with measurement of another anatomical structure to estimate gestational age of said fetus by approximating a distribution and uncertainty for each measurement, projecting each measurement distribution and uncertainty onto a gestational age axis of a gestational biometric chart to obtain a gestational age hypothesis for each measurement, estimating the distribution of the gestational age hypotheses, and fusing said age hypotheses using a variable bandwidth density-based fusion wherein a majority vote of said multiple hypotheses estimates said gestational age.

2. The method of claim 1, wherein said classifiers are trained to detect anatomical structures by providing a set of ellipses to represent contours of said anatomical structures, providing a database of pairs of contours and intensity sets that are represented by said contour, using a boosting technique with said database for training said classifiers to correctly detect contour/intensity-set pairs, but reject most non-contour/intensity-set pairs, and incorporating some non-contour/intensity-set pairs into said classifier, wherein said trained classifier outputs an image model that maps a set of image intensities to a likelihood of a presence of an anatomical structure.

3. The method of claim 2, wherein said classifiers explicitly incorporate translation, rotation, and scaling of said intensity-sets.

4. The method of claim 2, wherein said contour/intensity pairs in said database were selected by one or more experts in obstetrics.

5. The method of claim 1, wherein segmenting and measuring further comprise sweeping said 3-dimensional ultrasound fetal image wherein a plurality of planes are obtained by uniformly sampling the azimuthal and elevation angles;

and

for each plane of said plurality of planes, applying said contour classifiers to a plurality of locations, scales, orientations, and aspect ratios.

6. The method of claim 5, further comprising propagating said 2-dimensional contours to form a 3-dimensional shape of said anatomical structure.

7. The method of claim 2, wherein training said contour classifier comprises registering said contours to ellipses by learning global transformations, wherein global transformations include translation, scale, rotation, and aspect ratio, and applying an inverse global transformation.

8. The method of claim 7, wherein said ellipses include contours that deviate slightly from a perfect ellipse.

9. The method of claim 1, wherein said anatomical measurements include two or more measurements selected from the group including of fetal crown-rump length, head circumference, bi-parietal diameter, abdominal diameter, occipital-frontal diameter, and femur length.

10. A non-transitory program storage device readable by a computer, tangibly embodying a program of instructions executable by the computer to perform the method steps for segmenting and measuring anatomical structures in fetal ultrasound images, said method comprising the steps of:

providing a digitized 3-dimensional ultrasound image of a fetus comprising a plurality of intensities corresponding to a domain of points on a 3-dimensional grid;

providing a plurality of classifiers trained to detect anatomical structures in said image of said fetus;

segmenting and measuring an anatomical structure using said image classifiers by applying elliptical contour classifiers to said fetal ultrasound image, wherein a plurality of 2-dimensional contours characterizing said anatomical structure are detected; and

combining said anatomical structure measurement with measurement of another anatomical structure to estimate gestational age of said fetus by approximating a distribution and uncertainty for each measurement, projecting each measurement distribution and uncertainty onto a gestational age axis of a gestational biometric chart to obtain a gestational age hypothesis for each measurement, estimating the distribution of the gestational age hypotheses, and fusing said age hypotheses using a variable bandwidth density-based fusion wherein a majority vote of said multiple hypotheses estimates said gestational age.

11. The computer readable program storage device of claim 10, wherein said classifiers are trained to detect anatomical structures by providing a set of ellipses to represent contours of said anatomical structures, providing a database of pairs of contours and intensity sets that are represented by said contour, using a boosting technique with said database for training said classifiers to correctly detect contour/intensity-set pairs, but reject most non-contour/intensity-set pairs, and incorporating some non-contour/intensity-set pairs into said classifier, wherein said trained classifier outputs an image model that maps a set of image intensities to a likelihood of a presence of an anatomical structure.

12. The computer readable program storage device of claim 11, wherein training said contour classifier comprises registering said contours to ellipses by learning global transformations, wherein global transformations include translation, scale, rotation, and aspect ratio, and applying an inverse global transformation.

13. The computer readable program storage device of claim 12, wherein said ellipses include contours that deviate slightly from a perfect ellipse.

14. The computer readable program storage device of claim 11, wherein said classifiers explicitly incorporate translation, rotation, and scaling of said intensity-sets.

15. The computer readable program storage device of claim 11, wherein said contour/intensity pairs in said database were selected by one or more experts in obstetrics.

16. The computer readable program storage device of claim 10, wherein segmenting and measuring further comprise sweeping said 3-dimensional ultrasound fetal image wherein a plurality of planes are obtained by uniformly sampling the azimuthal and elevation angles;

and

for each plane of said plurality of planes, applying said contour classifiers to a plurality of locations, scales, orientations, and aspect ratios.

17. The computer readable program storage device of claim 16, the method further comprising propagating said 2-dimensional contours to form a 3-dimensional shape of said anatomical structure.

18. The computer readable program storage device of claim 10, wherein said anatomical measurements include two or more measurements selected from the group including of fetal crown-rump length, head circumference, bi-parietal diameter, abdominal diameter, occipital-frontal diameter, and femur length.