Method for providing a label of a body part on an x-ray image
One or more example embodiments relates to a computer-implemented a method for providing a label of a body part on an X-ray image, comprising receiving input data, wherein the input data is based on a red, green and blue (RGB) image of the body part, a depth image of the body part and an X-ray image of the body part; applying at least one trained function to the input data to generate output data, wherein the output data is the label of the body part, the label indicating a right body part or a left body part; and providing the output data.
1. A computer-implemented method for providing a label of a body part on an X-ray image, comprising:
receiving input data, the input data being based on at least two among a red, green and blue (RGB) image of the body part, a depth image of the body part or an X-ray image of the body part, the input data including,
context information or shape information derived from the RGB image, or
shape information derived from the depth image;
applying a plurality of trained functions to the input data to generate separate output data associated with the at least two among the RGB image, the depth image or the X-ray image, a total output data being provided by a consensus protocol based on the separate output data, the total output data being the label of the body part, and the label indicating a right body part or a left body part; and
providing the total output data.
2. The method according to claim 1 , wherein the input data comprises:
the context information or the shape information derived from the RGB image,
the shape information derived from the depth image, and
skeleton structure information derived from the X-ray image.
3. The method according to claim 1 , further comprising:
providing a manual confirmation notification or a manual correction notification.
4. The method according to claim 1 , wherein at least one among the plurality of trained functions is based on a machine learning algorithm.
5. The method according to claim 1 , wherein at least one among the plurality of trained functions is based on a rule-based model.
6. The method according to claim 1 , wherein the providing includes displaying the label as an overlay on the X-ray image.
7. The method according to claim 1 , wherein the body part is a paired body part.
8. A computer-implemented method for providing a plurality of trained functions for an X-ray system, comprising:
receiving input training data, the input training data being based on at least two among a red, green and blue (RGB) image of a body part, a depth image of the body part and an X-ray image of the body part, the input training data including,
context information or shape information derived from the RGB image, or
shape information derived from the depth image;
receiving separate output training data associated with the at least two among the RGB image, the depth image or the X-ray image, a total output training data being provided by a consensus protocol based on the separate output training data, the total output training data being a label of the body part, and the label indicating a right body part or a left body part;
training a plurality of trained functions based on the input training data and the separate output training data; and
providing the plurality of trained functions.
9. An X-ray system comprising:
a first interface configured to receive input data, the input data being based on at least two among a red, green and blue (RGB) image of a body part, a depth image of the body part or an X-ray image of the body part, the input data including,
context information or shape information derived from the RGB image, or
shape information derived from the depth image;
a computation unit configured to apply a plurality of trained functions to the input data to generate separate output data associated with the at least two among the RGB image, the depth image or the X-ray image, a total output data being provided by a consensus protocol based on the separate output data, the total output data being a label of the body part, and the label indicating a right body part or a left body part; and
a second interface configured to provide the total output data.
10. A non-transitory computer-readable medium comprising instructions which, when executed by a computer, cause the computer to carry out the method of claim 1 .
11. A non-transitory computer-readable medium comprising instructions which, when executed by a computer, cause the computer to carry out the method of claim 2 .
12. The method according to claim 2 , wherein the applying applies the plurality of trained functions to the input data to generate separate output data associated with each among the RGB image, the depth image and the X-ray image.
13. The method according to claim 2 , wherein the providing includes displaying the label as an overlay on the X-ray image.
14. The method according to claim 6 , wherein the providing includes displaying the label as an overlay on the X-ray image.
15. The method according to claim 1 , wherein
the body part is one among a pair of body parts, the pair of body parts including a right paired body part and a left paired body part; and
the plurality of trained functions are configured to determine whether a respective body part is the right paired body part or the left paired body part based on at least one of,
a bone position in the respective body part, or
a shape of the respective body part.
16. The method according to claim 1 , further comprising:
capturing the X-ray image using an X-ray detector when the body part is arranged in an interspace between an X-ray source and the X-ray detector; and
displaying the label with the X-ray image on a graphical user interface.
17. The method according to claim 1 , wherein the applying applies the plurality of trained functions to the input data to generate separate output data associated with each among the RGB image, the depth image and the X-ray image.
18. The method according to claim 1 , wherein
the plurality of trained functions includes a first trained function;
the applying applies the first trained function to the input data based on the RGB image; and
the first trained function is based on a machine learning algorithm.
19. The method according to claim 1 , wherein
the plurality of trained functions includes a first trained function;
the applying applies the first trained function to the input data based on the depth image; and
the first trained function is based on a machine learning algorithm.
20. The method according to claim 1 , wherein
the plurality of trained functions includes a first trained function;
the applying applies the first trained function to the input data based on the X-ray image; and
the first trained function is based on a rule-based model.
21. The method according to claim 1 , wherein
the plurality of trained functions includes a first trained function, a second trained function and a third trained function;
the applying applies
the first trained function to the input data based on the RGB image,
the second trained function to the input data based on the depth image, and
the third trained function to the input data based on the X-ray image;
each of the first trained function and the second trained function is based on a machine learning algorithm; and
the third trained function is based on a rule-based model.