Information search apparatus and method, and computer readable memory

• US 6,856,987 B2
• Filed: 09/05/2002
• Issued: 02/15/2005
• Est. Priority Date: 09/30/1998
• Status: Expired due to Fees

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Kobayashi, Yuji, Takata, Tomomi
• Primary Examiner(s): ALAM, SHAHID AL
• Application Number:
  US10/234,185
• Publication Number:
  US 20040002964A1
• Time in Patent Office: 894 Days
• Field of Search: 707/1, 707/2, 707/3, 707/4, 707/5, 707/104.1, 707/527, 707/104, 707/6, 707/8, 707/10, 715/527, 715/513, 706/46, 717/146, 382/224
• US Class Current: 1/1
• CPC Class Codes: G06F 16/48 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
• CPC Class Codes: G06F 16/5846 using extracted text
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99936 Pattern matching access
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing

Abstract

A search request input processing unit holds an input query word in a query word holding unit. An associative word unfolding processing unit acquires associative words in association with the query word held in the query word holding unit with reference to an associative word dictionary. An image content search processing unit using associative words makes a keyword search of image information on the basis of the obtained associative words and the query word. A sensory pattern unfolding processing unit obtains the query word and sensory patterns corresponding to associative words of the query word, and searches image information using feature amounts of the obtained sensory patterns. A search result integration processing unit integrates the search results obtained by the image content word search processing unit and sensory pattern search processing unit.

Claims

1. An information search apparatus for searching information based on an input query word, comprising:

first search means for determining a query keyword on the basis of the query word, and searching the information by comparing the query keyword with a feature keyword expressing a feature of the information, wherein said feature keyword is stored in association with the information;

second search means for determining a feature amount corresponding to the query word, and searching the information by comparing the determined feature amount with a feature amount extracted from the information, wherein said extracted feature amount is stored in association with the information; and

generation means for generating search result information based on both results obtained by said first and second search means.

2. The apparatus according to claim 1, further comprising:

an associative word dictionary in which index words and associative words associated with the index words are registered in correspondence with each other,

wherein said first search means determines as query keywords the query word and an associative word found by searching said associative word dictionary using the query word.

3. The apparatus according to claim 2, wherein said associative word dictionary registers an association strength indicating an association relationship strength between each associative word and index word,

said first search means obtains a search matching level of each information found by search on the basis of the association strength of each associative word, and

said second search means searches the information with reference to the search matching levels obtained by said first search means.

4. The apparatus according to claim 3, further comprising:

acquisition means for acquiring a contrary associative word indicating an antithetic concept to the query word, and

wherein said first search means lowers a search matching level of an image found by search using the contrary associative word acquired by said acquisition means.

5. The apparatus according to claim 2, further comprising:

a representative pattern dictionary for corresponding representative patterns in units of pairs of query word and associative words, and

wherein said second search means searches said representative pattern dictionary to acquire a corresponding pattern using a combination of a query word and associative word, and searches information on the basis of a feature amount of the acquired representative pattern.

6. The apparatus according to claim 1, wherein the information is image information, and the feature amount is a physical image feature amount obtained by analyzing the image information.

7. The apparatus according to claim 6, wherein the feature amount includes at least one of color scheme information, composition information, and shape information of an image.

8. An information search method for searching information based on an input query word, comprising:

a first search step for determining a query keyword on the basis of the query word, and searching the information by comparing the query keyword with a feature keyword expressing a feature of the information, wherein said feature keyword is stored in association with the information;

a second search step for determining a feature amount of a pattern corresponding to the query word, and searching the information by comparing the determined feature amount with a feature amount extracted from the information, wherein said extracted feature amount is stored in association with the information; and

generating search result information based on both results obtained in the first and second search steps.

9. The method according to claim 8, further comprising:

an associative word dictionary in which index words and associative words associated with the index words are registered in correspondence with each other,

wherein the first search step includes the step of determining as query keywords the query word and an associative word found by searching said associative word dictionary using the query word.

10. The method according to claim 9, wherein said associative word dictionary registers an association strength indicating an association relationship strength between each associative word and index word,

the first search step includes obtaining a search matching level of each information found by the search on the basis of the association strength of each associative word, and

the second search step searches the information with reference to the search matching levels obtained in the first search step.

11. The method according to claim 10, further comprising:

acquiring a contrary associative word indicating an antithetic concept to the query word, and

wherein the first search step includes lowering a search matching level of an image found by search using the contrary associative word acquired in the acquisition step.

12. The method according to claim 9, further comprising:

a representative pattern dictionary for corresponding representative patterns in units of pairs of query word and associative words, and

wherein the second search step includes searching said representative pattern dictionary to acquire a corresponding pattern using a combination of query word and associative word, and searching information on the basis of a feature amount of the acquired representative pattern.

13. The method according to claim 8, wherein the information is image information, and the feature amount is a physical image feature amount obtained by analyzing the image information.

14. The method according to claim 13, wherein the feature amount includes at least one of color scheme information, composition information, and shape information of an image.

15. A computer readable memory for storing a computer program for searching information based on an input query word, the computer program comprising:

code for a first search step to determine a query keyword on the basis of the query word, and to search information by comparing the query keyword with a feature keyword expressing a feature of the information, wherein said feature keyword is stored in association with the information;

code for a second search step to determine a feature amount corresponding to the query word, and to search information by comparing the determined feature amount with a feature amount extracted from the information, wherein said extracted feature amount is stored in association with the information; and

code to generate search result information based on both results obtained in the first and second search steps.