Granted Patent B2

US 12,724,968 · App. 18/521,035 · Granted Sep 1, 2026

Inventor: Shimpei Fujisaki (Tokyo, JP)

Assignee: CANON KABUSHIKI KAISHA

G06F40/295G06F16/164G06F40/284

Abstract

Claims (37)

1 . An information processing apparatus comprising:

a storage unit configured to store a plurality of named entities in association with information indicating whether to display the plurality of named entities;

one or more processors coupled to a memory that function as:

an extraction unit configured to extract a plurality of named entities by inputting document data to a named entity recognition model;

a determination unit configured to determine whether each of the extracted plurality of named entities represents a generic concept or a specific concept; and

a display control unit configured to display the extracted plurality of named entities in a relationship between the generic concept and the specific concept on a UI screen of a display device,

wherein the display control unit:

displays one of the extracted plurality of named entities representing the generic concept based on information on one of the stored plurality of named entities representing the generic concept, in a case where the one of the extracted plurality of named entities is similar to the one of the stored plurality of named entities; and

displays one of the extracted plurality of named entities representing the specific concept based on information on one of the stored plurality of named entities representing the specific concept, in a case where the one of the stored plurality of named entities is a named entity which is most similar to the one of the extracted plurality of named entities,

wherein the information indicating whether to display the plurality of named entities is initially set so that a named entity representing the generic concept is output and a named entity representing the specific concept is not output, the information being modifiable based on user instructions, and

wherein the similarity is calculated by using cosine similarity between an average of characteristic amount vectors of each token of the one of the extracted plurality of named entities and an average of characteristic amount vectors of each token of the one of the stored plurality of named entities.

2 . The information processing apparatus according to claim 1 , wherein the display control unit displays the plurality of the named entities by dividing into the determined generic concept or the determined specific concept.

3 . The information processing apparatus according to claim 2 , wherein, in a case where there are a generic concept level and a specific concept level in the plurality of the named entities determined to be the specific concept, the display control unit displays the named entity at the generic concept level at a position closer to the named entity determined to be the generic concept.

4 . The information processing apparatus according to claim 3 , wherein the extracted plurality of the named entities is stored in association with the information based on user instructions.

5 . The information processing apparatus according to claim 1 , wherein the display control unit:

calculates a similarity between each of the stored plurality of named entities representing the specific concept and each of the extracted plurality of named entities representing the specific concept.

6 . The information processing apparatus according to claim 1 , wherein the document data is generated by performing character recognition processing for a document.

7 . A method of controlling an information processing apparatus, comprising:

storing a plurality of named entities in association with information indicating whether to display the plurality of named entities;

extracting a plurality of named entities by inputting document data to a named entity recognition model;

determining whether each of the extracted plurality of named entities represents a generic concept or a specific concept; and

displaying the extracted plurality of named entities in a relationship between the generic concept and the specific concept on a UI screen of a display device in one group,

wherein in the displaying:

one of the extracted plurality of named entities representing the generic concept based on information on one of the stored plurality of named entities representing the generic concept is displayed on the UI screen, in a case where the one of the extracted plurality of named entities is most similar to the one of the stored plurality of named entities; and

one of the extracted plurality of named entities representing the specific concept based on information on one of the stored plurality of named entities representing the specific concept is displayed on the UI screen, in a case where the one of the stored plurality of named entities is a named entity which is most similar to the one of the extracted plurality of named entities,

wherein the information indicating whether to display the plurality of named entities is initially set so that a named entity representing the generic concept is output and a named entity representing the specific concept is not output, the information being modifiable based on user instructions, and

wherein the similarity is calculated by using cosine similarity between an average of characteristic amount vectors of each token of the one of the extracted plurality of named entities and an average of characteristic amount vectors of each token of the one of the stored plurality of named entities.

8 . A non-transitory computer readable storage medium storing a program executable by a computer to perform a method of controlling an information processing apparatus, the method comprising:

storing a plurality of named entities in association with information indicating whether to display the plurality of named entities;

extracting a plurality of named entities by inputting document data to a named entity recognition model;

determining whether each of the extracted plurality of named entities represents a generic concept or a specific concept; and

displaying the extracted plurality of named entities in a relationship between the generic concept and the specific concept on a UI screen of a display device in one group,

wherein in the displaying:

one of the extracted plurality of named entities representing the generic concept based on information on one of the stored plurality of named entities representing the generic concept is displayed on the UI screen, in a case where the one of the extracted plurality of named entities is most similar to the one of the stored plurality of named entities; and

one of the extracted plurality of named entities representing the specific concept based on information on one of the stored plurality of named entities representing the specific concept is displayed on the UI screen, in a case where the one of the stored plurality of named entities is a named entity which is most similar to the one of the extracted plurality of named entities,

wherein the information indicating whether to display the plurality of named entities is initially set so that a named entity representing the generic concept is output and a named entity representing the specific concept is not output, the information being modifiable based on user instructions, and

wherein the similarity is calculated by using cosine similarity between an average of characteristic amount vectors of each token of the one of the extracted plurality of named entities and an average of characteristic amount vectors of each token of the one of the stored plurality of named entities.

Assignments (1)

Priority Claims (1)

Continuity (1)

References Cited (18)