Age estimation apparatus, age estimation method, and age estimation program
View Patent ↗Provided are an age estimation apparatus, an age estimation method, and an age estimation program capable of obtaining a recognition result closely matching the result perceived by human. An age estimation apparatus 10 for estimating an age of a person on image data includes a dimension compressor 11 for applying dimension compression to the image data to output low dimensional data; and an identification device 12 for estimating an age of a person on the basis of a learning result using a feature amount contained in the low dimensional data, wherein a parameter used for the dimension compression by the dimension compressor 11 and the feature amount used for age estimation by the identification device 12 are set on the basis of a result of an evaluation of a generalization capability using a weighting function that shows a degree of seriousness of an age estimation error for every age, and learning of the identification device 12 is performed on the basis of the weighting function.
1. An age estimation apparatus for estimating an age of an object contained by an image data, the age estimation apparatus comprising:
a processor configured to control any of a dimension compression unit and an identification unit;
the dimension compression unit configured to output low dimensional data of image data by applying a dimension compression to the image data by extracting a feature from the image data; and
the identification unit configured to minimize a value of a predetermined function, to set a learned parameter for determining an age bracket in which an error of an estimated age is minimized, and to estimate the age using the learned parameter, wherein
the age is estimated from a feature which is contained in the low dimensional data,
the error is calculated between the estimated age and a correct age,
the predetermined function is defined by the error,
the learned parameter is set by a weighting function which is a function of the age bracket,
the correct age being the age of the object contained by the image data,
the weighting function is defined on the basis of a standard deviation of each of a plurality of perceptual ages, and
the identification unit is further configured to estimate the age to closely match a human recognition result.
2. The age estimation apparatus according to claim 1 , wherein a plurality of the weighting functions are used.
3. An age estimation method for estimating an age of an object contained by an image data, comprising:
outputting, to an identification device, low dimensional data of the image data by applying a dimension compression to the image data by extracting a feature from the image data;
estimating an age of the object on the image data from a feature extracted from the image data contained in the low dimensional data;
defining a value of a predetermined function by an error between the estimated age and a correct age;
minimizing the value of the predetermined function;
setting a learned parameter for determining an age bracket by a weighting function which is a function of age for determining the age bracket in which the error of the estimated age is minimized, wherein
the estimated age is estimated from the feature contained in the low dimensional data using the learned parameter,
the correct age being the age of the object contained by the image data,
defining the weighting function on the basis of a standard deviation of each of a plurality of perceptual ages, and
estimating the age to closely match a human recognition result.
4. The age estimation method according to claim 3 , wherein a plurality of the weighting functions are used.