Height estimation method, height estimation apparatus, and program

A height estimation apparatus comprising:

• a processor;
• and
• a storage medium having computer program instructions stored thereon, when executed by the processor, perform to:
• extract, from a first input image in which a subject image that is an image of a subject is captured, a feature point coordinate that is a coordinate of a feature point predetermined along a skeleton of the subject image;
• estimate a coordinate of a first subject frame that is a frame surrounding the subject image in the first input image;
• derive a height of the first subject frame in the first input image in accordance with the coordinates of the first subject frame, and generate a distance addition pattern that is an addition pattern of a distance between a feature point coordinate and another feature point coordinate and a correction coefficient for an individual missing pattern that is a pattern of a combination of one or a plurality of the feature point coordinates that are not extracted among the plurality of the feature point coordinates predetermined;
• extract a feature point coordinate from a second input image in which an object image that is an image of an object and the subject image are captured;
• estimate a coordinate of a second subject frame that is a frame surrounding the subject image in the second input image and estimate a coordinate of an object frame that is a frame surrounding the object image in the second input image;
• select the individual missing pattern and the correction coefficient in accordance with the feature point coordinate extracted from the second input image;
• select an object height that is a height of the object in the object image in the second input image in accordance with information on the object image in the second input image;
• and
• add up a distance between a feature point coordinate and another feature point coordinate extracted from the second input image in accordance with the missing pattern selected and derive an estimated value of a height of the subject in accordance with a result of adding up the distance between the feature point coordinate and the other feature point coordinate in the second input image, the correction coefficient selected, the object height, and the coordinate of the object frame.