Learning model, storage medium storing diagnostic program, ultrasonic diagnostic apparatus, ultrasonic diagnostic system, image diagnostic apparatus, machine learning apparatus, learning data creation apparatus, learning data creation method, and storage medium storing learning data creation program

An ultrasonic diagnostic apparatus comprising:

• an ultrasonic probe that transmits and receives ultrasonic waves to and from a subject;
• a hardware processor;
• and
• a learning model, wherein the learning model is machine-learned using learning data formed with a pair of:
• first ultrasonic image data based on a reception signal for image generation received by the ultrasonic probe from an object;
• and
• second correct answer data,
• wherein the second correct answer data is obtained by generating second ultrasonic image data by performing coordinate transformation on the first ultrasonic image data, obtaining a first correct answer by input from a user viewing the second ultrasonic image data, and acquiring the second correct answer data from an inverse transformation of the first correct answer data;
• wherein the hardware processor executes the learning model by:
• obtaining a third ultrasonic image data based on a reception signal that is received by the ultrasonic probe and is intermediate image data that has not been subjected to processing including coordinate transformation;
• and
• causes the learning model to output a first inference result from the third ultrasonic image data,
• wherein the first inference result includes a probability distribution image representing the probability of being a range of a structure, subjects the probability distribution image to the coordinate transformation, and calculates a characteristic value of the structure from the probability distribution image after the coordinate transformation.

An ultrasonic diagnostic system comprising:

• an ultrasonic probe that transmits and receives ultrasonic waves to and from a subject;
• a hardware processor;
• and
• a learning model, wherein the learning model is machine-learned using learning data formed with a pair of:
• first ultrasonic image data based on a reception signal for image generation received by the ultrasonic probe from an object;
• and
• second correct answer data,
• wherein the second correct answer is obtained by generating second ultrasonic image data by performing coordinate transformation on the first ultrasonic image data, obtaining a first correct answer by input from a user viewing the second ultrasonic image data, and acquiring the second correct answer data from an inverse transformation of the first correct answer data;
• wherein the hardware processor executes the learning model by:
• obtaining a third ultrasonic image data being based on a reception signal that is received by the ultrasonic probe and is intermediate image data that has not been subjected to processing including coordinate transformation, and
• causing the learning model to output a first inference result from the third ultrasonic image data,
• wherein the first inference result includes a probability distribution image representing the probability of being a range of a structure, subjects the probability distribution image to the coordinate transformation, and
• calculates a characteristic value of the structure from the probability distribution image after the coordinate transformation.