Patent attributes
An object-motion detection method includes extracting, from each of a first image representing a detection-target object and generated by an imaging unit and a second image representing the detection-target object and generated after the first image by the imaging unit, an object area including the detection-target object, obtaining, for each of the first image and the second image, a reference point indicating a boundary between a moving part of the detection-target object and a fixed part of the detection-target object, the moving part moving when the detection-target object performs a certain motion, the fixed part not moving compared to the moving part when the detection-target object performs the certain motion, and obtaining, for each of the first image and the second image, a position of the moving part of the object area in a first area locating on a side including the moving part with respect to the reference point.