| 摘要 |
<P>PROBLEM TO BE SOLVED: To efficiently and automatically learn an image processing algorithm by a genetic algorithm. <P>SOLUTION: A learning device 1 includes: a designated data input unit 15 for accepting an input of designated data; a candidate data extraction unit 12 for extracting candidate data representing position data in learning image data 11; a comparison/evaluation unit 14 for, when an image processing group is assumed to be gene, calculating a gene evaluation function that gives a higher evaluation as the candidate data is located closer to the designated data; and an automatic adjustment unit 13 that, by referring to the gene evaluation value that the comparison/evaluation unit 14 has calculated, changes the generation of the gene cluster by the genetic algorithm and outputs excellent genes as learning results. <P>COPYRIGHT: (C)2013,JPO&INPIT |
