SYSTEM AND METHOD FOR DIFFERENTIATING A TISSUE OF INTEREST FROM ANOTHER PART OF A MEDICAL SCANNER IMAGE
One or more example embodiments provides a system and a method for differentiating a tissue of interest from another part of a medical scanner image, in particular pectoral muscle tissue from breast tissue in an X-ray mammography image. The method comprises providing a medical scanner image; inputting input data into a trained artificial neural network, the input data being based on the provided medical scanner image; generating, by the trained artificial neural network, output data based on the input data, the output data indicating a one-dimensional borderline between at least a part of the tissue of interest and the at least one other part of the medical scanner image; and outputting an output signal comprising or based on the generated output data.
1 . A system for differentiating a tissue of interest from at least one other part of a medical scanner image, the system comprising:
an input module configured to receive a medical scanner image;
a computing device configured to implement a trained artificial neural network and configured to generate output data based on the medical scanner image, the output data indicating a one-dimensional borderline between at least a part of the tissue of interest and at least one other part within the medical scanner image; and
an output module configured to output an output signal comprising or based on the output data.
2 . The system of claim 1 , wherein the output data indicates the one-dimensional borderline by at least one of:
polynomial parameters,
spline parameters, or
line parameters.
3 . The system of claim 1 , wherein the output data indicates the one-dimensional borderline using a Fast Fourier Transformation.
4 . The system of claim 1 , wherein
the computing device is further configured to implement a mask-generating module,
the mask-generating module is configured to generate a segmentation mask based on the output data, and
the output module is configured to output the segmentation mask.
5 . The system of claim 1 , wherein
the computing device is configured to implement a resizing module,
the resizing module is configured to reduce a size of the medical scanner image to obtain a reduced-size medical scanner image, and
the trained artificial neural network is configured to receive the reduced-size medical scanner image as an input.
6 . The system of claim 1 , wherein the medical scanner image is an X-ray mammography image and the tissue of interest is pectoral muscle tissue.
7 . The system of claim 1 , wherein
the computing device is further configured to implement an orientation determining module,
the orientation determining module is configured to determine, based on the borderline indicated by the output data, at least one of an orientation of the tissue of interest within the medical scanner image, a position of the tissue of interest within the medical scanner image, or a parameter of the medical scanner image, and
the output signal comprises data indicating at least one of the determined orientation, position or parameter.
8 . A computer-implemented method for differentiating a tissue of interest from at least one other part of a medical scanner image, the method comprising:
providing a medical scanner image;
inputting input data into a trained artificial neural network, the input data being based on the provided medical scanner image;
generating, by the trained artificial neural network, output data based on the input data, the output data indicating a one-dimensional borderline between at least a part of the tissue of interest and the at least one other part of the medical scanner image; and
outputting an output signal comprising or based on the generated output data.
9 . The method of claim 8 , wherein the output data indicates the one-dimensional borderline by at least one of:
polynomial parameters,
spline parameters, or
line parameters.
10 . The method of claim 8 , wherein the output data indicates the one-dimensional borderline using a Fast Fourier Transformation.
11 . The method of claim 8 , further comprising:
generating a segmentation mask based on the output data, wherein the output signal comprises the generated segmentation mask.
12 . A computer-implemented method for training an artificial neural network for differentiating a tissue of interest from at least one other part of a medical scanner image, the method comprising:
providing a set of medical scanner training images;
providing, for each medical scanner training image, label data indicating a labeled one-dimensional borderline between at least a part of the tissue of interest and the at least one other part of the medical scanner training image;
providing an artificial neural network configured to receive at least one of the medical scanner training images as input data and to generate output data based on the input data, the output data indicating a generated one-dimensional borderline between at least the part of the tissue of interest and the at least one other part of the medical scanner training image; and
training the artificial neural network with supervised learning using a loss function penalizing differences between the output data and the label data.
13 . The method of claim 12 , wherein the providing the label data for the set of medical scanner training images comprises:
providing, for each medical scanner training image, MSTI, a ground truth segmentation mask, the ground truth segmentation mask indicating the tissue of interest or indicating a remainder of the medical scanner training image;
performing binary erosion on the ground truth segmentation mask to create an eroded segmentation mask; and
subtracting the eroded segmentation mask from the ground truth segmentation mask to generate the label data.
14 . A non-transitory computer program product comprising executable program code configured to, when executed by a system, cause the system to perform the method of claim 8 .
15 . A non-transient computer-readable data storage medium comprising executable program code configured to, when executed by a system, cause the system to perform the method of claim 8 .
16 . A non-transient computer-readable data storage medium comprising executable program code configured to, when executed by a system, cause the system to perform the method of claim 12 .
17 . The system of claim 2 , wherein
the computing device is configured to implement a resizing module,
the resizing module is configured to reduce a size of the medical scanner image to obtain a reduced-size medical scanner image, and
the trained artificial neural network is configured to receive the reduced-size medical scanner image as an input.
18 . The system of claim 2 , wherein
the computing device is further configured to implement an orientation determining module,
the orientation determining module is configured to determine, based on the borderline indicated by the output data, at least one of an orientation of the tissue of interest within the medical scanner image, a position of the tissue of interest within the medical scanner image, or a parameter of the medical scanner image, and
the output signal comprises data indicating at least one of the determined orientation, position or parameter.
19 . The system of claim 4 , wherein
the computing device is configured to implement a resizing module,
the resizing module is configured to reduce a size of the medical scanner image to obtain a reduced-size medical scanner image, and
the trained artificial neural network is configured to receive the reduced-size medical scanner image as an input.
20 . The system of claim 4 , wherein
the computing device is further configured to implement an orientation determining module,
the orientation determining module is configured to determine, based on the borderline indicated by the output data, at least one of an orientation of the tissue of interest within the medical scanner image, a position of the tissue of interest within the medical scanner image, or a parameter of the medical scanner image, and
the output signal comprises data indicating at least one of the determined orientation, position or parameter.