Method, program and apparatus for storing document and/or image using invariant values calculated from feature points and method, program and apparatus for retrieving document based on stored document and/or image

• US 8,036,497 B2
• Filed: 02/15/2006
• Issued: 10/11/2011
• Est. Priority Date: 03/01/2005
• Status: Expired due to Fees

• Current Assignee: Osaka Prefcture University
• Original Assignee: Osaka Prefcture University
• Inventors: Kise, Koichi, Iwamura, Masakazu, Nakai, Tomohiro
• Primary Examiner(s): Repko; Jason M
• Assistant Examiner(s): Thomas; Mia M
• Application Number:
  US11/885,567
• Publication Number:
  US 20080177764A1
• Time in Patent Office: 2,064 Days
• Field of Search: 382/190, 382/305, 382/306, 707/E17.008, 358/1.9
• US Class Current: 382/305
• CPC Class Codes: G06F 16/583 using metadata automatically derived from the content
• CPC Class Codes: G06V 30/40 Document-oriented image-based pattern recognition
• CPC Class Codes: G06V 30/418 Document matching, e.g. of document images

Abstract

A document/image retrieval method for retrieving a document/image corresponding to a captured digital image from a database by comparing features calculated based on feature points of the captured digital image with features preliminarily calculated based on feature points of each of documents and/or images stored in the database, the method comprising: extracting the feature points from the captured digital image; defining a local set of feature points for each of the extracted feature points; selecting feature points from the defined local set to define a feature point subset of the local set; determining invariant values as values characterizing the defined subset for combinations of the feature points in the subset, the invariant values being invariant to a geometric transformation; calculating a feature by combining the determined invariant values; and performing a voting process on the documents and/or images in the database based on the preliminarily calculated features of the documents and/or images; whereby the document/image corresponding to the captured digital image is retrieved from the database.

Claims

1. A document and/or image retrieval method for retrieving a document and/or image corresponding to a captured digital image from a database by comparing features calculated based on feature points of the captured digital image with features calculated based on feature points of each of documents and/or images stored in the database, the method comprising:

extracting the feature points from the captured digital image;

defining a local set of feature points for each of the extracted feature points;

selecting subsets of feature points from the defined local set;

determining invariant values as values characterizing each selected subset for combinations of the feature points in the subset, the invariant values being invariant to a geometric transformation;

calculating a feature by combining the determined invariant values; and

performing a voting process on the documents and/or images in the database based on the preliminarily calculated features of the documents and/or images;

so that the document and/or image corresponding to the captured digital image is retrieved from the database.

2. A document and/or image retrieval method according to claim 1, wherein the invariant values are cross-ratios.

3. A document and/or image retrieval method according to claim 1, wherein the invariant values are invariant to an affine transformation.

4. A document and/or image retrieval method according to claim 1, wherein the invariant values are invariant to a similarity transformation.

5. A document and/or image storage method, which causes a computer to execute the steps of:

inputting a document and/or image;

assigning an ID to the input document and/or image;

extracting feature points defining an image arrangement from the input document and/or image; and

performing a predetermined process on each of the extracted feature points;

the predetermined process comprising the steps of;

(1) selecting n feature points which are nearest to a feature point p of interest; and

(2) performing a predetermined process on each of all possible sets of m feature points, m<

n, selected from the selected n feature points;

the predetermined process in the performing step comprising the steps of;

(a) determining features, for each of which includes a set of invariant values, for all possible sets of d points, wherein d is a number not greater than a predetermined number m, selected from an m-point set of interest;

(b) determining an index of a hash table based on the determined features through a predetermined computation; and

(c) storing the features in relation to a point ID and a document ID in the hash table, the features being determined in the features determining step using the determined hash index, the point ID being assigned to the feature point p and the document ID being assigned to the document and/or image from which the feature point p is extracted.

6. A document and/or image retrieval method for retrieving a document and/or image stored by a storage method as recited in claim 5, the retrieval method causing a computer to execute the steps of:

reading a captured image;

extracting feature points defining an image arrangement from the read image;

performing a predetermined process on each of the extracted feature points;

the predetermined process comprising the steps of;

(1) selecting n feature points which are nearest to a feature point p of interest; and

(2) performing a predetermined process on each of all possible sets of m feature point, m<

n, selected from the selected n feature points;

the predetermined process in the performing step comprising the steps of;

(a) determining features for all possible sets of d points, wherein d is a number not greater than a predetermined number m selected from an m-point set of interest;

(b) determining an index of a hash table based on the determined features through a predetermined computation; and

(c) acquiring features of a preliminarily input document and/or image from the hash table based on the determined hash index, comparing the features determined in the features determining step with the acquired features, and voting for a document ID having matching features;

and after the selecting step and the performing steps, specifying a document ID of a document and/or image which matches the captured image based on a voting result.

7. A non transitory computer readable medium storing a document and/or image storage program, which causes a computer to execute the steps of:

inputting a document and/or image;

assigning an ID to the input document and/or image;

extracting feature points defining an image arrangement from the input document and/or image; and

performing a predetermined process on each of the extracted feature points;

the predetermined process including the steps of;

(1) selecting n feature points which are nearest to a feature point p of interest; and

(2) performing a predetermined process on each of all possible sets of m feature points, m<

n, selected from the selected n feature points;

the predetermined process in the performing step including the steps of;

(a) determining features for each of which includes a set of invariant values for all possible sets of d points wherein d is a number not greater than a predetermined number m, selected from an m-point set of interest;

(b) determining an index of a hash table based on the determined features through a predetermined computation; and

(c) storing the features in relation to a point ID and a document ID in the hash table, the features being determined in the features determining step using the determined hash index, the point ID being assigned to the feature point p and the document ID being assigned to the document and/or image from which the feature point p is extracted.

8. A non transitory computer readable medium storing a document and/or image retrieval program for retrieving a document and/or image input employing the storage program as recited in claim 7, the retrieval program causing a computer to execute the steps of:

reading a captured image;

extracting feature points defining an image arrangement from the read image;

performing a predetermined process on each of the extracted feature points from the read image;

the predetermined process including the steps of;

(1) selecting n feature points which are nearest to a feature point p of interest; and

(2) performing a predetermined process on each of all possible sets of m feature points, m<

n, selected from the selected n feature points;

the predetermined process in the performing step including the steps of;

(a) determining features for all possible sets of d points, wherein d is a number not greater than a predetermined number m, selected from an m-point set;

(b) determining an index of a hash table based on the determined features through a predetermined computation; and

(c) acquiring features of a preliminarily input document and/or image from the hash table based on the determined hash index, comparing the features determined in the features determining step with the acquired features,

and voting for a document ID having matching features; and

after the selecting step and the performing steps, specifying a document ID of a document/image which matches the captured image based on a voting result.

9. A document and/or image storage apparatus comprising:

a processor;

an input section which inputs a document and/or image;

a feature point extracting section which extracts feature points defining an image arrangement from the input document and/or image;

a feature point selecting section which selects n feature points nearest to an extracted feature point p of interest; and

a feature storing section which performs a predetermined process on each of all possible sets of m feature points, m<

n, selected from the selected n feature points;

the predetermined process including the steps of;

(a) determining features for each of which includes a set of invariant values for all possible sets of d points, wherein d is a number not greater than a predetermined number m, selected from an m-point set of interest;

(b) determining an index of a hash table based on the determined features through a predetermined computation; and

(c) storing the features in relation to a point ID and a document ID in the hash table, the features being determined in the features determining step using the determined hash index, the point ID being assigned to the feature point p and the document ID being assigned to the document and/or image from which the feature point p is extracted.

10. A document and/or image retrieval apparatus comprising:

a processor;

a reading section which reads a captured image;

a feature point extracting section which extracts feature points defining an image arrangement from the read image;

a feature point selecting section which selects n feature points nearest to an extracted feature point p of interest; and

a voting section which performs a predetermined process on each of all possible sets of m feature points, m<

n, selected from the selected n feature points;

the predetermined process including the steps of;

(a) determining features for all possible sets of d points, wherein d is a number not greater than a predetermined number m, selected from an m-point set of interest;

(b) determining an index of a hash table based on the determined features through a predetermined computation; and

(c) acquiring features of a preliminarily input document and/or image from the hash table based on the determined hash index, comparing the features determined in the features determining step with the acquired features, and voting for a document ID having matching features;

and a document ID specifying section which specifies a document ID of a document/image which matches the captured image based on a voting result determined by votes corresponding to the respective feature points.