Medical image data processing technique

1 . A method for processing medical image data of a patient's body, the method comprising:

(a) obtaining medical image data of a patient's body;

(b) identifying a first portion of the medical image data that represents at least one predefined anatomical region of the patient's body;

(c) determining, based on the first portion of the medical image data, a first criterion for allocating pixels or voxels of the medical image data to a first surface represented by the medical image data;

(d) obtaining a second criterion for allocating pixels or voxels of the medical image data to a second surface represented by the medical image data, the second criterion determined based on the medical image data and on at least one of:

(i) a histogram analysis of values of pixels or voxels of the medical image data; and

(ii) a trained machine learning model

(e) generating, based on the second criterion and the medical image data, a second surface model of the second surface; and

(f) identifying a second portion of the second surface as modeled by the second surface model that represents the at least one predefined anatomical region, wherein

the first criterion is determined based on values of pixels or voxels of one or more sets of pixels or voxels of the medical image data, and

the pixels or voxels of each set lie along a same straight line associated with the respective set, wherein the line has a predefined spatial relationship to the second surface as modeled by the second surface model.

2 . The method of claim 1 , wherein the first criterion comprises a value of pixels or voxels of the medical image data representing a part of the first surface.

3 . The method of claim 2 , wherein the value is representative of a strength of a signal received by a detector of a medical imaging device.

4 . The method of claim 1 , further comprising generating, based on the first criterion and the medical image data, a first surface model of the first surface.

5 . The method of claim 1 , wherein the second criterion is predefined, user-defined or determined based on the medical image data.

6 . The method of claim 5 , wherein at least one of the following conditions is fulfilled:

(i) the first portion of the medical image data is identified based on the second portion; and

(ii) the first criterion is determined based on the second portion.

7 . The method of claim 1 , wherein at least one of the following conditions is fulfilled:

the line intersects the second portion;

the line has a predefined angle relative to at least a part of the second portion;

the line intersects the second portion at an intersection point and is orthogonal to the second portion at the intersection point;

the line has a predefined angle relative to another line associated with another one of the one or more sets; and

the line has a predefined distance relative to another line associated with another one of the one or more sets.

8 . The method of claim 1 , wherein the first criterion is determined based on a trained machine learning model and/or a gradient analysis of pixels or voxels of at least one of the sets.

9 . The method of claim 1 , wherein multiple first portions of the medical image data are identified and wherein multiple first criteria are determined based on the multiple first portions for allocating pixels or voxels of the medical image data to respective first surfaces represented by the medical image data.

10 . The method of claim 1 , further comprising:

adapting the medical image data before performing at least one of the method steps, wherein, during the at least one method step, the adapted medical image data is used instead of the medical image data, wherein adapting the medical image data comprises at least one adaption selected from:

(i) excluding imaging artifacts from the medical image data;

(ii) excluding one or more clinical objects represented by the medical image data; and

(iii) changing a spatial orientation of the medical image data.

11 . The method of claim 10 , wherein the step of adapting the medical image data comprises:

(i) performing, iteratively, two or more of the at least one adaption, or

(ii) performing, iteratively, one of the at least one adaption multiple times, to gradually adapt the medical image data.

12 . The method of claim 1 , wherein each surface is of a surface type, and wherein the surface type is one of:

(i) a bone surface of the patient's body,

(ii) a skin surface of the patient's body, or

(iii) a surface of a clinical object.

13 . The method of claim 12 , wherein at least the step (c) is performed multiple times, each time for a different surface type of the first surface.

14 . A computing system comprising at least one processor configured to:

(i) obtain medical image data of a patient's body;

(ii) identify a first portion of the medical image data that represents at least one predefined anatomical region of the patient's body;

(iii) determine, based on the first portion of the medical image data, a first criterion for allocating pixels or voxels of the medical image data to a first surface represented by the medical image data;

(iv) obtain a second criterion for allocating pixels or voxels of the medical image data to a second surface represented by the medical image data, the second criterion determined based on the medical image data and on at least one of:

(a) a histogram analysis of values of pixels or voxels of the medical image data; and

(b) a trained machine learning model

(v) generate, based on the second criterion and the medical image data, a second surface model of the second surface; and

(vi) identify a second portion of the second surface as modeled by the second surface model that represents the at least one predefined anatomical region, wherein

the first criterion is determined based on values of pixels or voxels of one or more sets of pixels or voxels of the medical image data, and

the pixels or voxels of each set lie along a same straight line associated with the respective set, wherein the line has a predefined spatial relationship to the second surface as modeled by the second surface model.

15 . A computer readable storage medium storing a computer program comprising instructions which, when executed on at least one processor, cause the at least one processor to:

(i) obtain medical image data of a patient's body;

(ii) identify a first portion of the medical image data that represents at least one predefined anatomical region of the patient's body;

(iii) determine, based on the first portion of the medical image data, a first criterion for allocating pixels or voxels of the medical image data to a first surface represented by the medical image data

(iv) obtain a second criterion for allocating pixels or voxels of the medical image data to a second surface represented by the medical image data, the second criterion determined based on the medical image data and on at least one of:

(a) a histogram analysis of values of pixels or voxels of the medical image data;

(b) a trained machine learning model and

(v) generate, based on the second criterion and the medical image data, a second surface model of the second surface; and

(vi) identify a second portion of the second surface as modeled by the second surface model that represents the at least one predefined anatomical region, wherein

the first criterion is determined based on values of pixels or voxels of one or more sets of pixels or voxels of the medical image data, and

the pixels or voxels of each set lie along a same straight line associated with the respective set, wherein the line has a predefined spatial relationship to the second surface as modeled by the second surface model.

16 . The method of claim 13 , further comprising:

adapting the medical image data after performing at least step (c), wherein adapting the medical image data comprises excluding one or more clinical objects based on the surface type.