摘要 |
<p><P>PROBLEM TO BE SOLVED: To improve an identification performance of sensory data. <P>SOLUTION: An image/brain signal storage section 20 stores a plurality of pairs of brain activity data representing brain wave signals that are measured when image data is displayed, at respective measuring points of the brain of a subject and the image data. An image feature amount learning section 22 performs a kernel canonical correlation analysis on the plurality of the pairs of brain activity data and image data stored in the image/brain signal storage section 20 and learns a weight in a canonical correlation variable represented by a linear sum of a kernel function such that a correlation of the brain activity data and the image data is increased. The canonical correlation variable in the image side using the learned weight is defined as an image feature amount. <P>COPYRIGHT: (C)2012,JPO&INPIT</p> |