Computer-readable recording medium storing information processing program, information processing method, and information processing device

An information processing method implemented by a computer to execute processing comprising:

• acquiring a video image in which a first person appears;
• identifying, for each body part of two body parts that form a pair of left and right body parts among body parts of the first person in the video image, positional information indicating a position of the each body part by analyzing the acquired video image;
• generating a first component that indicates a sum of two pieces of the identified positional information for the two body parts;
• generating a third component that indicates an absolute value of a second component that indicates a difference in the two pieces of the identified positional information for the two body parts;
• and
• training a model that outputs posture information about a second person from a video image in which the second person appears, based on training data in which the generated first component and the generated third component are associated with a label indicating a correct action of the first person.

An information processing apparatus comprising:

• memory;
• and
• processor circuitry coupled to the memory, the processor circuitry being configured to perform processing including:
• acquiring a video image in which a first person appears;
• identifying, for each body part of two body parts that form a pair of left and right body parts among body parts of the first person in the video image, positional information indicating a position of the each body part by analyzing the acquired video image;
• generating a first component that indicates a sum of two pieces of the identified positional information for the two body parts;
• generating a third component that indicates an absolute value of a second component that indicates a difference in the two pieces of the identified positional information for the two body parts;
• and
• training a model that outputs posture information about a second person from a video image in which the second person appears, based on training data in which the generated first component and the generated third component are associated with a label indicating a correct action of the first person.