| US 7,519,226 B2 | ||
| Form search apparatus and method | ||
| Kitahiro Kaneda, Kanagawa (Japan); Ken-ichi Kazumi, Kanagawa (Japan); and Tomotoshi Kanatsu, Tokyo (Japan) | ||
| Assigned to Canon Kabushiki Kaisha, Tokyo (Japan) | ||
| Filed on May 12, 2005, as Appl. No. 11/127,188. | ||
| Application 11/127188 is a division of application No. 09/571411, filed on May 15, 2000, granted, now 6,970,601. | ||
| Claims priority of application No. 1999-133071 (JP), filed on May 13, 1999; application No. 2000-019148 (JP), filed on Jan. 27, 2000; and application No. 2000-138449 (JP), filed on May 11, 2000. | ||
| Prior Publication US 2005/0238257 A1, Oct. 27, 2005 | ||
| Int. Cl. G06K 9/62 (2006.01) | ||
| U.S. Cl. 382—224 [382/190; 382/209; 382/278] | 48 Claims |
| 1. A method of classifying forms comprising one or more processor implementing the steps of:
registering a plurality of template forms with features of table attribute areas included in each template form, the features
including information of the position and size of each table area;
inputting image data representative of an input form;
segmenting the input image data into a plurality of areas, each area having one of a plurality of attributes, wherein the
plurality of attributes includes a table attribute, a frame attribute, and a figure attribute;
extracting features of the plurality of areas segmented in said segmenting step;
performing an approximate selection, by selecting some template forms from among the plurality of registered template fonts
according to a result of comparison between the features extracted from the segmented table attribute areas of the input form
and the features of table attribute areas of the registered template forms, wherein the compared features include the numbers
of areas;
calculating each score of some template forms selected in said approximate selection step by comparing the extracted features
of the plurality of segmented areas of the input form with the features of table attribute areas of the some template forms
selected in said approximate selection step, wherein the compared features of the input form include information of the position
and size of the areas having the table attribute or the frame attribute or the figure attribute, and wherein the compared
features of the some template forms are information of the position and size of the areas having the table attribute;
making a further selection by selecting a temp1ate form matching the input form from among the some template forms selected in said approximate selecting step based on the
scores calculated in said calculating step; and
outputting an output corresponding to the template form selected in said further selection step.
|