PROBLEM TO BE SOLVED: To provide an attribute estimation device for estimating an attribute division a little by little than in a conventional manner without increasing the number of images used for learning.SOLUTION: An attribute estimation device includes: image acquisition means 110 for acquiring an image relating to an object for estimating an attribute feature amount calculation means 121 for acquiring an attribute feature amount from the image a first age range likelihood calculation means 122 for acquiring an attribute likelihood representing a degree by which the acquired attribute feature amount includes each feature of the identification divisions second age range likelihood calculation means 123 generation likelihood calculation means 124 and integration estimation means 125 for acquiring an overlap division likelihood representing a degree of having the feature of the overlap division in which the attribute feature amount overlaps with the overlap division from among the identification divisions, and estimating the attribute division to which the object belongs by using an overlap division likelihood. When the age of a person is selected as the attribute of the estimation target, an age estimation device 100 can be realized.SELECTED DRAWING: Figure 3
