Thickness calculation method, calculation device, and program

1 . A thickness calculating method for calculating a thickness of a target tissue in a target living body by a computer that includes a processor or a plurality of processors, the thickness calculating method comprising:

the processor or the plurality of processors performing:

a data acquisition step that acquires received data based on a reception signal output when an ultrasonic wave transmitted from an ultrasonic probe is reflected by the target tissue within the target living body and is received by the ultrasonic probe;

a living body attribute classification step that classifies the target living body into one of a first attribute relating to a physique of the target living body and a second attribute relating to the physique based on the acquired received data; and

a thickness calculating step of (i) when the target living body is classified into the first attribute in the living body attribute classification step, calculating the thickness of the target tissue at a predetermined part of the target living body by inputting the acquired received data into a first training model that learned in advance and (ii) when the target living body is classified into the second attribute in the living body attribute classification step, calculating the thickness of the target tissue by inputting the acquired received data into a second training model that learned in advance, wherein

the first training model learned in advance using, as teaching data, first sample data, the first sample data being data for a biological sample having the first attribute, and a thickness label related to a thickness of a tissue associated with the first sample data,

the second training model learned in advance using, as teaching data, second sample data, the second sample data being data for a biological sample having the second attribute and a thickness label associated with the second sample data,

each of the thickness label associated with the first sample data and the thickness label associated with the second sample data is a label that identifies, by segmentation of the acquired received data, positions of a first end and a second end in a thickness direction of the target tissue,

the segmentation is a three-class segmentation that uses the label to indicate whether each pixel in the acquired received data represents one of a fat layer, a muscle layer, and an internal organ,

the label specifies a boundary between the fat layer and the muscle layer and represents the first end and the second end, and

in the thickness calculating step, the thickness of the target tissue is calculated using the positions of the first end and the second end, wherein the positions are output by input of the acquired received data.

2 . The thickness calculating method according to claim 1 , wherein the first attribute and the second attribute are attributes classified according to a magnitude of a parameter related to the physique.

3 . The thickness calculating method according to claim 2 , wherein the parameter represents at least one of a fat thickness, a body fat percentage, or a Body Mass Index (BMI).

4 . The thickness calculation method according to claim 1 , wherein

in the living body attribute classification step, the target living body is classified into one of the first attribute or the second attribute by using a third training model different from the first training model and the second training model.

5 . The thickness calculation method according to claim 4 , wherein

the first attribute and the second attribute are attributes classified according to magnitude of a parameter related to the physique, and

the third training model learned in advance using, as teaching data, the first sample data and a physique label related to a parameter associated with the first sample data, and the second sample data and a physique label associated with the second sample data.

6 . A calculation device for calculating a thickness of a target tissue in a target living body, the calculation device comprising:

a storage device for storing a first training model that was trained in advance and a second training model that was trained in advance; and

a processor or a plurality of processors, wherein the processor or the plurality of processors performs:

a data acquisition step that acquires received data based on a reception signal output when an ultrasonic wave transmitted from an ultrasonic probe is reflected by the target tissue within the target living body and is received by the ultrasonic probe;

a living body attribute classification step that classifies the target living body into one of a first attribute relating to a physique of the target living body and a second attribute relating to the physique based on the acquired received data; and

a thickness calculating step of (i) when the target living body is classified into the first attribute in the living body attribute classification step, calculating the thickness of the target tissue at a predetermined part of the target living body by inputting the acquired received data into the first training model and (ii) when the target living body is classified into the second attribute in the living body attribute classification step, calculating the thickness of the target tissue by inputting the acquired received data into the second training model, wherein

the first training model learned in advance using, as teaching data, first sample data, the first sample data being data for a biological sample having the first attribute, and a label related to a thickness of a tissue associated with the first sample data,

the second training model learned in advance using, as teaching data, second sample data, the second sample data being data for a biological sample having the second attribute and a label associated with the second sample data,

each of the label associated with the first sample data and the label associated with the second sample data identifies, by segmentation of the acquired received data, positions of a first end and a second end in a thickness direction of the target tissue,

the segmentation is a three-class segmentation that uses the label to indicate whether each pixel in the acquired received data represents one of a fat layer, a muscle layer, and an internal organ,

the label specifies a boundary between the fat layer and the muscle layer and represents the first end and the second end, and

in the thickness calculating step, the thickness of the target tissue is calculated using the positions of the first end and the second end, wherein the positions are output by input of the acquired received data.

7 . The calculation device according to claim 6 , further comprising:

the ultrasonic probe that transmits ultrasonic waves to the target living body and that outputs the acquired received data.

8 . A non-transitory computer-readable storage medium storing a program, the program causing a computer to realize a function of calculating a thickness of a target tissue in a target living body, the program causing the computer to execute operations comprising:

executing a data acquisition step of acquiring received data based on a reception signal output when an ultrasonic wave transmitted from an ultrasonic probe is reflected by the target tissue within the target living body and is received by the ultrasonic probe;

executing a living body attribute classification step of classifying the target living body into one of a first attribute relating to a physique of the target living body and a second attribute relating to the physique based on the acquired received data; and

executing a thickness calculating step of (i) when the target living body is classified into the first attribute in the living body attribute classification step, calculating the thickness of the target tissue at a predetermined part of the target living body by inputting the acquired received data into a first training model that learned in advance and (ii) when the target living body is classified into the second attribute in the living body attribute classification step, calculating the thickness of the target tissue by inputting the acquired received data into a second training model that learned in advance, wherein

the first training model learned in advance using, as teaching data, first sample data, the first sample data being data for a biological sample having the first attribute, and a label related to a thickness of a tissue associated with the first sample data,

the second training model learned in advance using, as teaching data, second sample data, the second sample data being data for a biological sample having the second attribute and a label associated with the second sample data,

each of the label associated with the first sample data and the label associated with the second sample data identifies, by segmentation of the acquired received data, positions of a first end and a second end in a thickness direction of the target tissue,

the segmentation is a three-class segmentation that uses the label to indicate whether each pixel in the acquired received data represents one of a fat layer, a muscle layer, and an internal organ,

the label specifies a boundary between the fat layer and the muscle layer and represents the first end and the second end, and

in the thickness calculating step, the thickness of the target tissue is calculated using the positions of the first end and the second end, wherein the positions are output by input of the acquired received data.