Posture estimation device, posture estimation method, and posture estimation program

摘要

A candidate region extraction unit (110) extracts candidate regions from an image by using a common characteristic amount. A part region extraction unit (120) separates a part region, which has a high likelihood of being a part, from a second-part candidate region other than the part region. A characteristic amount calculation unit (130) calculates the intrinsic characteristic amount of each of the part region and the second-part candidate region. A characteristic amount correction unit (140) corrects the intrinsic characteristic amount of either one of the candidate regions on the basis of the length, the width, and the angle of the other candidate region. A determination unit (150) determines whether or not the part region and the second-part candidate region are parts that constitute a pair on the basis of the similarity between the corrected intrinsic characteristic amount and the intrinsic characteristic amount of the other candidate region.

主权项

1. A posture estimation apparatus that estimates a paired part based on an image including a whole or part of an articulated object having the paired part, the posture estimation apparatus comprising a processor configured to:
extract a plurality of candidate regions of the paired part from the image using a common feature value indicating features of a shape common to an indefinite number of articulated objects regarding the paired part; extract a first candidate region having a maximum likelihood of being the paired part from the plurality of candidate regions and designates a region other than the first candidate region as a second candidate region; calculate a unique feature value indicating a feature unique to the articulated object to be estimated about the paired part in each of the first candidate region and the second candidate region; correct a unique feature value of one of the first candidate region and the second candidate region based on at least one of a length, width, and angle of the other candidate region; and determine whether or not the first candidate region and the second candidate region are the paired parts based on similarity between the corrected unique feature value and a unique feature value of the other candidate region.