ATTRIBUTE IDENTIFICATION APPARATUS, ATTRIBUTE IDENTIFICATION MODEL LEARNING APPARATUS, METHOD AND PROGRAM

To identify an attribute of an image with high accuracy.SOLUTION: An attribute identification unit 230 identifies at least one attribute the image has on the basis of a multilayer neural network model including a convolution neural network with an image whose attribute is unknown as input, which outputs the probability that the image has each of a plurality of attributes and a layer outputting a probability having each of the plurality of attributes according to a correlation between the attributes, using output of the convolution neural network and a potential function indicating the correlation between the attributes.SELECTED DRAWING: Figure 5