Method of and system for extracting predetermined elements from input document based upon model which is adaptively modified according to variable amount in the input document

• US 6,456,738 B1
• Filed: 07/15/1999
• Issued: 09/24/2002
• Est. Priority Date: 07/16/1998
• Status: Expired due to Fees

• Current Assignee: Ricoh Company Limited
• Original Assignee: Ricoh Company Limited
• Inventors: Tsukasa, Kohchi
• Primary Examiner(s): COUSO, YON JUNG
• Application Number:
  US09/353,929
• Time in Patent Office: 1,167 Days
• Field of Search: 382/175, 382/181, 382/155, 382/156, 382/190, 382/209, 382/218, 382/180, 707/505, 707/511, 707/516, 707/500, 707/517
• US Class Current: 382/175
• CPC Class Codes: G06V 30/40 Document-oriented image-based pattern recognition

Abstract

Elements are extracted from an input document image according to a predetermined extraction model. Based upon the variability in a predetermined set of layout characteristics of the extracted elements, the extraction model is adaptively modified to improve future performance in extracting the elements.

Claims

1. A method of extracting one or more elements from a document using model data, the model data including at least a template, comprising:

a) determining the template for a predetermined document type, the template having a set of predetermined characteristics for each of the elements;

b) inputting at least one input document;

c) extracting the elements having the predetermined characteristics from the input document according to the model data;

d) storing the extracted characteristics of the elements in the model data;

e) determining a distance value between the stored characteristics and a corresponding one in the model data;

f) determining a variable amount based upon the distance value for each of the element; and

g) modifying the model data based upon the variable amount.

2. The method of extracting an element of claim 1 wherein the model data is selected from a plurality of sets of the model data.

3. The method of extracting an element of claim 2 wherein the variable amount is determined for each of the characteristics for each of the elements based upon a comparison of the characteristic of the model data and the corresponding stored characteristics of the input document which has the closest characteristics to that of the model data.

4. The method of extracting an element of claim 3 wherein the characteristic in the model data is modified based the variable amount of the corresponding one of the stored extracted characteristics.

5. The method of extracting an element of claim 4 wherein the characteristic is modified if the variable amount is above a first predetermined threshold value.

6. The method of extracting an element of claim 4 wherein the model data is modified by adding a new template if the variable amount is above a second predetermined threshold value.

7. The method of extracting an element of claim 4 wherein the model data additionally includes a noise reduction potential index (NI) for each of the elements.

8. The method of extracting an element of claim 1 wherein the characteristics includes a coordinate, a font size and a font type.

9. The method of extracting an element of claim 4 wherein a predetermined set of extraction parameters in the model data is modified based upon an extraction confidence value and an adjacent area similarity value if the variable amount is below a third predetermined threshold value.

10. A system for extracting one or more elements from a document using model data, comprising:

a model generation unit for generating model data for a predetermined document type, the model data including at least a template, the template having a set of predetermined characteristics for each of the elements;

a document input unit for inputting at least one input document;

an element extraction unit connected to said model generator and said document input unit for extracting the elements having the predetermined characteristics from the input document according to the model data;

a characteristics storage unit connected to said element extraction unit for storing the extracted characteristics of the elements in the model data;

a learning process unit connected to said characteristics storage unit for determining a distance value between the stored characteristics and a corresponding one in the model data and for determining a variable amount based upon the distance value; and

a model updating unit connected to said learning process unit and said model generation unit for modifying the model data based upon the variable amount.

11. The system for extracting an element of claim 10 wherein said model generation unit generates a plurality of sets of the model data and further comprising a model selection unit connected to said model generation unit for selecting one of the sets of the model data.

12. The system for extracting an element of claim 11 wherein said learning process unit determines the variable amount for each of the characteristics for each of the elements based upon a comparison of the characteristic of the model data and the corresponding stored characteristics of the input document which has the closest characteristics to that of the model data.

13. The system for extracting an element of claim 12 wherein said model updating unit modifies the characteristic in the template based the variable amount.

14. The system for extracting an element of claim 13 wherein said model updating unit modifies the characteristic if the variable amount is above a first predetermined threshold value.

15. The system for extracting an element of claim 13 wherein said model updating unit modifies the model data by adding a new template if the variable amount is above a second predetermined threshold value.

16. The system for extracting an element of claim 13 wherein the model data additionally includes a noise reduction potential index (NI) for each of the elements.

17. The system for extracting an element of claim 10 wherein the characteristics includes a coordinate, a font size and a font type.

18. The system for extracting an element of claim 13 wherein the model data includes a predetermined set of extraction parameters, the extraction parameter being modified based upon an extraction confidence value and an adjacent area similarity value if the variable amount is below a third predetermined threshold value.