Bone age estimation method and apparatus
Disclosed are a bone age estimation method and a bone age estimation apparatus. The bone age estimation method may comprise the steps of: extracting a region of interest including a cervical spine region from a lateral cephalometric radiographic image obtained by imaging a subject's cervical spine, by using a first deep learning model; extracting landmarks from the extracted region of interest by using a second deep learning model; calculating a landmark numerical value on the basis of the extracted landmarks; and providing maturity information of a maturation stage of the cervical spine on the basis of the calculated landmark numerical value.
1. A bone age estimation method, comprising the steps of:
extracting a region of interest including a cervical spine region from a lateral cephalometric radiographic image obtained by imaging a subject's cervical spine, by using a first deep learning model;
extracting landmarks from the extracted region of interest by using a second deep learning model;
displaying landmarks as points in the region of interest on the basis of a user input selecting landmarks to be displayed and landmarks to not be displayed from the extracted landmarks;
calculating a landmark numerical value on the basis of the displayed landmarks; and
providing maturity information of a maturation stage of the cervical spine on the basis of the calculated landmark numerical value,
wherein the step of providing maturity information includes:
a step of visually providing the maturity information of the cervical spine on the basis of a gender of the subject, age-specific standardization information for a concavity of a fourth cervical spine lower border corresponding to the gender of the subject, and a concavity ratio of the fourth cervical spine lower border of the subject, and
a step of providing a graph corresponding to the age-specific standardization information and a graph corresponding to the concavity ratio of the fourth cervical spine lower border of the subject to be overlaid.
2. The bone age estimation method of claim 1 , wherein the step of calculating a landmark numerical value includes:
a step of calculating a landmark numerical value including at least one of a ratio of a height to a length of the cervical spine lower border, a ratio of a height to a length of the cervical spine lower border that draws a vertical line, and a concavity ratio of the cervical spine lower border, on the basis of the coordinates of the landmarks.
3. The bone age estimation method of claim 1 , wherein the step of extracting landmarks includes:
a step of changing a coordinate of the extracted landmarks on the basis of the user input when a user input to change the extracted landmark is received.
4. The bone age estimation method of claim 3 , further comprising:
a step of training the second deep learning model on the basis of the changed coordinate when the coordinate of the extracted landmark is changed by the user input.
5. The bone age estimation method of claim 1 , further comprising:
a step of adjusting a brightness of the region of interest and reversing left and right of the region of interest after extracting the region of interest.
6. The bone age estimation method of claim 1 , wherein the step of extracting landmarks includes:
a step of providing a numerical value obtained by calculating a progress degree that the landmarks are extracted by a percentage, on the basis of a predetermined period.
7. The bone age estimation method of claim 1 , wherein the region of interest includes at least one of a second cervical spine, a third cervical spine, and a fourth cervical spine of the subject.
8. The bone age estimation method of claim 7 , wherein the landmarks are points obtained by anatomically measuring and marking at least one of the second cervical spine, the third cervical spine, and the fourth cervical spine of the subject.
9. The bone age estimation method of claim 1 , wherein each of the extracted landmarks is assigned with a tracking number on the basis of a position of each of the landmarks, and the step of displaying landmarks further includes displaying tracking numbers corresponding to the displayed landmarks in the region of interest.
10. A non-transitory computer readable storage medium storing instructions that are operable with a processor to execute the method of claim 1 .
11. A bone age estimation apparatus which performs a bone age estimation method, the apparatus comprising:
a memory; and
a processor;
wherein the memory stores instructions executable by the processor that when executed by the processor, the processor controls the bone age estimation apparatus to allow the bone age estimation apparatus to extract a region of interest including a cervical spine region from a lateral cephalometric radiographic image obtained by imaging a subject's cervical spine, by using a first deep learning model, extract landmarks from the extracted region of interest by using a second deep learning model, display landmarks as points in the region of interest on the basis of a user input selecting landmarks to be displayed and landmarks to not be displayed from the extracted landmarks, calculate a landmark numerical value on the basis of the displayed landmarks, and provide maturity information of a maturation stage of the cervical spine on the basis of the calculated landmark numerical value,
wherein the processor controls the bone age estimation apparatus to allow the bone age estimation apparatus to visually provide the maturity information of the cervical spine on the basis of a gender of the subject, age-specific standardization information for a concavity of a fourth cervical spine lower border corresponding to the gender of the subject, and a concavity ratio of the fourth cervical spine lower border of the subject,
wherein the processor controls the bone age estimation apparatus to allow the bone age estimation apparatus to provide a graph corresponding to the age-specific standardization information and a graph corresponding to the concavity ratio of the fourth cervical spine lower border of the subject to be overlaid.
12. The bone age estimation apparatus of claim 11 , wherein the processor controls the bone age estimation apparatus to allow the bone age estimation apparatus to calculate a landmark numerical value including at least one of a ratio of a height to a length of the cervical spine lower border, a ratio of a height to a length of the cervical spine lower border that draws a vertical line, and a concavity ratio of the cervical spine lower border, on the basis of the coordinates of the landmarks.
13. The bone age estimation apparatus of claim 11 , wherein the processor controls the bone age estimation apparatus to allow the bone age estimation apparatus to change a coordinate of the extracted landmarks on the basis of the user input when a user input to change the extracted landmark is received.
14. The bone age estimation apparatus of claim 13 , wherein the processor controls the bone age estimation apparatus to allow the bone age estimation apparatus to train the second deep learning model on the basis of the changed coordinate when the coordinate of the extracted landmark is changed by the user input.
15. The bone age estimation apparatus of claim 11 , wherein the landmarks are points obtained by anatomically measuring and marking at least one of a second cervical spine, a third cervical spine, and a fourth cervical spine of the subject.