Ovarian follicle count and size determination using transvaginal ultrasound scans

1. A computer-implemented method for determining an ovarian follicle count and size, comprising:

receiving a pair of 2-D (two-dimensional) transvaginal ultrasound scans, each scan comprising a plurality of 2-D ultrasound images captured along a translation direction of an ultrasound probe used to capture said scan, the respective translation directions being at least approximately orthogonal, and for each of said scans, the method further comprising:

determining at least one of ovarian rotation and truncation in a detected ovary of each of the 2-D (two-dimensional) transvaginal ultrasound scans;

detecting candidate follicles in the plurality of 2-D ultrasound images of the 2-D transvaginal ultrasound scan, wherein detecting candidate follicles in the plurality of 2-D ultrasound images comprises:

approximating a volume image from the plurality of 2-D ultrasound images;

detecting the ovary in the approximated volume image;

detecting possible candidate follicles within the detected ovary;

determining a volume for each candidate follicle; and

rejecting the candidate follicles that have a determined volume below a defined volume threshold for subsequent segmenting;

segmenting the detected candidate follicles to determine an average diameter for each candidate follicle;

selecting the 2-D transvaginal ultrasound scan including the candidate follicle having the largest average diameter; and

presenting the determined average diameters and the segmentations of the detected candidate follicles of the selected 2-D transvaginal ultrasound scan.

2. The computer-implemented method of claim 1 , wherein:

said selecting the 2-D transvaginal ultrasound scan including the follicle having the largest average diameter is based on the determined average diameters.

3. The computer-implemented method of claim 1 , wherein detecting candidate follicles in the plurality of 2-D ultrasound images further comprises labelling the detected candidate follicles for further processing.

4. The computer-implemented method of claim 1 , wherein segmenting the detected candidate follicles to determine an average diameter for each candidate follicle within the detected ovary further comprises:

defining a seed point for each candidate follicle in the plurality of 2-D ultrasound images, and for each seed point, applying an adaptive region growing algorithm to the plurality of 2-D ultrasound images to segment the candidate follicle associated with said seed point.

5. The computer-implemented method of claim 4 , further comprising:

receiving a user-defined additional seed point and applying the adaptive region growing algorithm to the plurality of 2-D ultrasound images to segment the candidate follicle associated with said user-defined additional seed point; and/or

receiving a user-identified existing seed point and deleting the user-identified existing seed point from the defined seed points.

6. The computer-implemented method of claim 4 , further comprising:

determining a form factor of a pixel area obtained with the adaptive region growing algorithm; and

rejecting the candidate follicle if said form factor is below a defined form factor threshold.

7. The computer-implemented method of claim 3 , further comprising:

determining an area of overlap between overlapping segmented candidate follicles; and

deleting the overlapping segmented candidate follicle having the smaller average diameter if said area of overlap exceeds a defined overlap threshold.

8. The computer-implemented method of claim 1 , wherein determining an average diameter for each candidate follicle comprises:

determining an average diameter of a follicle portion in each 2-D ultrasound image comprising a portion of the follicle;

selecting the 2-D ultrasound image having the largest average diameter of the follicle portion; and

selecting said largest average diameter as the average diameter of the follicle.

9. A computer program product comprising a non-transitory computer readable storage medium having computer readable program instructions embodied therewith for, when executed on a processor arrangement of an ultrasound image processing device, cause said processor arrangement to implement a method comprising:

receiving a pair of 2-D (two-dimensional) transvaginal ultrasound scans, each scan comprising a plurality of 2-D ultrasound images captured along a translation direction of an ultrasound probe used to capture said scan, the respective translation directions being at least approximately orthogonal, and for each of said scans, the method further comprising:

determining at least one of ovarian rotation and truncation in a detected ovary of each of the 2-D (two-dimensional) transvaginal ultrasound scans;

detecting candidate follicles in the plurality of 2-D ultrasound images of the 2-D transvaginal ultrasound scan wherein detecting candidate follicles in the plurality of 2-D ultrasound images comprises:

approximating a volume image from the plurality of 2-D ultrasound images;

detecting the ovary in the approximated volume image;

detecting possible candidate follicles within the detected ovary;

determining a volume for each candidate follicle; and

rejecting the candidate follicles that have a determined volume below a defined volume threshold for subsequent segmenting;

segmenting the detected candidate follicles to determine an average diameter for each candidate follicle;

selecting the 2-D transvaginal ultrasound scan including the candidate follicle having the largest average diameter;

presenting the determined average diameters and the segmentations of the detected candidate follicles of the selected 2-D transvaginal ultrasound scan.

10. An ultrasound image processing device comprising a processor arrangement adapted to implement a method comprising:

receiving a pair of 2-D (two-dimensional) transvaginal ultrasound scans, each scan comprising a plurality of 2-D ultrasound images captured along a translation direction of an ultrasound probe used to capture said scan, the respective translation directions being at least approximately orthogonal, and for each of said scans, the method further comprising:

determining at least one of ovarian rotation and truncation in a detected ovary of each of the 2-D (two-dimensional) transvaginal ultrasound scans;

detecting candidate follicles in the plurality of 2-D ultrasound images of the 2-D transvaginal ultrasound scan, wherein detecting candidate follicles in the plurality of 2-D ultrasound images comprises:

approximating a volume image from the plurality of 2-D ultrasound images;

detecting the ovary in the approximated volume image;

detecting possible candidate follicles within the detected ovary;

determining a volume for each candidate follicle; and

rejecting the candidate follicles that have a determined volume below a defined volume threshold for subsequent segmenting;

segmenting the detected candidate follicles to determine an average diameter for each candidate follicle;

selecting the 2-D transvaginal ultrasound scan including the candidate follicle having the largest average diameter; and

presenting the determined average diameters and the segmentations of the detected candidate follicles of the selected 2-D transvaginal ultrasound scan

wherein:

the ultrasound image processing device further comprises a computer program product of the processor arrangement being adapted to execute computer readable program instructions embodied by said computer program product.

11. An ultrasound imaging system comprising:

an ultrasound image processing device comprising: a processor arrangement adapted to implement a method comprising:

receiving a pair of 2-D (two-dimensional) transvaginal ultrasound scans, each scan comprising a plurality of 2-D ultrasound images captured along a translation direction of an ultrasound probe used to capture said scan, the respective translation directions being at least approximately orthogonal, and for each of said scans, the method further comprising:

determining at least one of ovarian rotation and truncation in a detected ovary of each of the 2-D (two-dimensional) transvaginal ultrasound scans;

detecting candidate follicles in the plurality of 2-D ultrasound images of the 2-D transvaginal ultrasound scan, wherein detecting candidate follicles in the plurality of 2-D ultrasound images comprises:

approximating a volume image from the plurality of 2-D ultrasound images;

detecting the ovary in the approximated volume image;

detecting possible candidate follicles within the detected ovary;

determining a volume for each candidate follicle; and

rejecting the candidate follicles that have a determined volume below a defined volume threshold for subsequent segmenting;

segmenting the detected candidate follicles to determine an average diameter for each candidate follicle;

selecting the 2-D transvaginal ultrasound scan including the candidate follicle having the largest average diameter; and

presenting the determined average diameters and the segmentations of the detected candidate follicles of the selected 2-D transvaginal ultrasound scan; and

an ultrasound probe for providing the ultrasound image processing device with said pair of transvaginal ultrasound scans.