# US 2012/0254191 A1 · Method and system for concept sumarization

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2012/0254191 A1
- Document ID: US20120254191A1
- Status: Published application
- Kind: A1
- Application number: 13/077,995
- Filed: 2011-04-01
- Published: 2012-10-04
- Earliest claimed: 2011-04-01
- Assignees: Yahoo! Inc.
- Inventors: Subhajit SANYAL; Dhruv Kumar Mahajan; Sundararajan Sellamanickam
- Web page: https://patentgrep.com/patent/US20120254191A1

## CPC

- G06F 16/58: Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A method and a system for summarizing a concept are provided. A query corresponding to a concept is received from a user. A plurality of images and corresponding descriptive information may be collected based on the query. The plurality of images and the descriptive information may be processed to form feature vectors and processed descriptive information respectively. Further, one or more topics may be identified for the plurality of images. Each of the plurality of images may be assigned with one or more topic distribution values corresponding to the one or more topics. The one or more topics correspond to the processed descriptive information. A sparse set of images may be determined based on the feature vectors and the assigned topic distribution values, to summarize the concept. Also, a target summary may be built from the summarized concept, by regularizing one or more distribution constraints.

## Claims

### Claim 1

A system for summarizing a concept for a summarization size, the system comprising:
a receiving module to receive a query corresponding to the concept;
a concept collecting module to collect a plurality of images along with descriptive information corresponding thereto, the plurality of images being collected from a source corresponding to the query;
a processing module comprising:
an image processing module configured to convert the plurality of images into corresponding feature vectors, and
a text processing module configured to process the descriptive information corresponding to each image of the plurality of images;
a topic distribution module configured to identify one or more topics for the plurality of images based on the processed descriptive information, and assign one or more topic distribution values corresponding to the one or more topics to each image of the plurality of images;
a summarizing module to define a sparse kernel classifier to determine a sparse set of images from the plurality of images to summarize the concept, the sparse set of images being determined based on the feature vectors and the assigned topic distribution values; and
an output module to provide the sparse set of images responsive to the query.

### Claim 2 (depends on 1)

The system of claim 1, wherein the receiving module further receives the source for collecting the plurality of images.

### Claim 3 (depends on 1)

The system of claim 1, wherein the concept collecting module comprising at least one adapter to collect the plurality of images and the descriptive information corresponding thereto from the source, wherein the at least one adapter extracts the descriptive information for each of the plurality of images.

### Claim 4 (depends on 1)

The system of claim 1 further comprising a database for storing the processed descriptive information.

### Claim 5 (depends on 1)

The system of claim 1, wherein the topic distribution module identifies the one or more topics by performing a latent semantic analysis of the processed descriptive information.

### Claim 6 (depends on 1)

The system of claim 1 further comprising a regularization module to facilitate regularization of one or more distribution constraints for building a target summary from the summarized concept, the one or more distribution constraints being regularized by utilizing a regularization term to minimize divergence between the summarized concept and the target summary.

### Claim 7 (depends on 1)

The system of claim 1, wherein the topic distribution module is further configured to determine the one or more topic distribution values of the one or more topics corresponding to each image, each of the topic distribution values comprising probability value for assigning a corresponding topic to an image of the plurality of images.

### Claim 8 (depends on 1)

The system of claim 1, wherein the sparse set of images being determined based on the summarization size.

### Claim 9 (depends on 1)

The system of claim 1, wherein the output module displays the summarized concept to a user.

### Claim 10

A system for summarizing a concept, the system comprising:
a concept collecting module to collect a plurality of images along with descriptive information corresponding thereto, the plurality of images and the descriptive information corresponding to the concept;
a processing module comprising:
an image processing module configured to convert the plurality of images into corresponding feature vectors, and
a text processing module configured to process the descriptive information corresponding to each image;
a topic distribution module configured to identify one or more topics for the plurality of images based on the processed descriptive information, and assign topic distribution values of the one or more topics to each image of the plurality of images;
a summarizing module configured to define a sparse kernel classifier to determine a sparse set of images from the plurality of images to summarize the concept, the sparse set of images being determined based on the feature vectors, and the assigned topic distribution values; and
a regularization module for facilitating regularization of one or more distribution constraints for building a target summary from the summarized concept.

### Claim 11 (depends on 10)

The system of claim 10 further comprising a receiving module configured to receive a query from a user, the query corresponding to the concept.

### Claim 12 (depends on 10)

The system of claim 10, wherein the plurality of images and the corresponding descriptive information being collected from a source corresponding to the concept.

### Claim 13 (depends on 10)

The system of claim 10, wherein the topic distribution module identifies the one or more topics by performing a latent semantic analysis of the processed descriptive information.

### Claim 14 (depends on 10)

The system of claim 10, wherein the topic distribution module is further configured to determine the one or more topic distribution values of the one or more topics corresponding to each image, each of the topic distribution values comprising probability value for assigning a corresponding topic to an image of the plurality of images.

### Claim 15 (depends on 10)

The system of claim 10, wherein the one or more distribution constraints received from a user.

### Claim 16 (depends on 10)

The system of claim 10, wherein the one or more distribution constraints being regularized by utilizing a regularization term to minimize divergence between the summarized concept and the target summary.

### Claim 17 (depends on 10)

The system of claim 10, wherein the sparse set of images being determined based on a summarization size.

### Claim 18 (depends on 10)

The system of claim 10 further comprising an output module to provide the target summary to a user.

### Claim 19

A computer-implemented method for summarizing a concept, the method comprising:
receiving a query corresponding to the concept;
collecting a plurality of images along with descriptive information corresponding thereto;
forming feature vectors with processed descriptive information by processing the plurality of images and the descriptive information;
assigning each of the plurality of images with one or more topic distribution values corresponding to one or more topics, the one or more topics corresponding to the processed descriptive information;
determining a sparse set of images from the plurality of images to summarize the concept, the sparse set of images being determined based on the feature vectors and the assigned topic distribution values; and
providing the sparse set of images responsive to the query.

### Claim 20 (depends on 19)

The method of claim 19 further comprising receiving information of the source for collecting the plurality of images therefrom.

### Claim 21 (depends on 19)

The method of claim 19 further comprising identifying the one or more topics for the plurality of images based on the processed descriptive information.

### Claim 22 (depends on 19)

The method of claim 19 further comprising determining the one or more topic distribution values of the one or more topics corresponding to each image, each of the topic distribution values comprising probability value for assigning the corresponding topic to an image of the plurality of images.

### Claim 23 (depends on 19)

The method of claim 19 further comprising regularizing one or more distribution constraints for building a target summary from the summarized concept, the one or more distribution constraints being regularized by utilizing a regularization term to minimize divergence between the summarized concept and the target summary.

### Claim 24 (depends on 19)

The method of claim 19, wherein the sparse set of images is determined by defining a sparse kernel classifier.

### Claim 25

A computer program product stored on a non-transitory computer-readable medium that when executed by a processor, performs a method for summarizing a concept, the method comprising:
receiving a query corresponding to the concept;
collecting a plurality of images along with descriptive information corresponding thereto, the plurality of images being collected from a source corresponding to the query;
processing the plurality of images and the descriptive information to form feature vectors and processed descriptive information respectively, each of the plurality of images corresponding to one of the feature vectors;
assigning each of the plurality of images with one or more topic distribution values corresponding to one or more topics, the one or more topics corresponding to the processed descriptive information;
determining a sparse set of images from the plurality of images to summarize the concept, the sparse set of images being determined based on the feature vectors and the assigned topic distribution values; and
providing the sparse set of images responsive to the query.
