| 摘要 |
PROBLEM TO BE SOLVED: To automatically determine a falsification detection threshold when falsification detection is performed. SOLUTION: A falsification detection part divides an area into uniform detection units, and calculates variation information of difference information of each divided area included in the detection units for each cell of each detection unit. It is understood that a cell 901, a cell 902, and a cell 903 have values of 8, 170, and 140, respectively. After calculating the variation information in this way, the falsification detection part calculates a threshold of the difference information for determining a falsification area in an image which is scanned and read using calculated variation of the difference information. In an area without falsification, both of an average value and standard deviation are comparatively small, and in an area with the falsification, both of the average value and the standard deviation are comparatively large. Namely, if the information is effectively used, a binarized threshold of the difference information is obtained. COPYRIGHT: (C)2011,JPO&INPIT
|
