Automated slide comparator

Patent metadata

Field	Value
Publication number	US-9792527-B2
Application number	US-201514882630-A
Country	US
Kind code	B2
Filing date	Oct 14, 2015
Priority date	Oct 14, 2015
Publication date	Oct 17, 2017
Grant date	Oct 17, 2017

Abstract

In response to an identification of an input slide, aspects generate text content confidence scores that represent amounts of similarity of slide text content of the input slide to compared text content of a slides that are included within slide presentation files of a repository. Graphic element content confidence scores are generated that represent amounts of similarity of graphic content of the input slide to compared graphic element content of the respective ones of the repository slides. Similarity confidence scores are generated for respective ones of the repository slides as functions of weighted averages of the generated text content confidence scores and graphic element content confidence scores, wherein the graphic element content confidence scores are weighted differently from the text content confidence scores. The repository slides may be ranked for similarity to the input slide as a function of the generated similarity confidence scores.

First claim

What is claimed is:

1. A computer-implemented method for automated slide content comparison, the method comprising executing on a computer processor the steps of:
in response to an identification of an input slide, generating text content confidence scores that represent amounts of similarity of slide text content of the input slide to compared text content of a plurality of slides that are included within slide presentation files of a repository;
generating graphic element content confidence scores that represent amounts of similarity of graphic content of the input slide to compared graphic element content of the respective ones of the plurality of slides;
generating similarity confidence scores for each of the respective ones of the plurality of slides as functions of weighted averages of the text content confidence scores and the graphic element content confidence scores generated for the respective ones of the plurality of slides, wherein the graphic element content confidence scores are weighted differently from the text content confidence scores; and
ranking the plurality of slides for similarity to the input slide as a function of the generated similarity confidence scores.

2. The method of claim 1 , wherein the step of generating the similarity confidence scores for each of the respective ones of the plurality of slides as the functions of the weighted averages of the text content confidence scores and the graphic element content confidence scores generated for the respective ones of the plurality of slides comprises:
generating first weighted averages of the graphic element content confidence scores and the text content confidence scores for each of the plurality of slides of the each slides as functions of a first differential weighting of the graphic element content confidence scores relative to the text content confidence scores;
comparing the graphic element content confidence scores of the plurality of slides to an image content confidence threshold value that indicates a strength of match of an attribute of the graphic content of the input slide to a corresponding attribute of the graphic content of the plurality of slides;
for each of the plurality of slides having a compared graphic element content confidence score that meets the image content confidence threshold value, generating second weighted averages of the graphic element content confidence scores and the text content confidence scores as functions of a second differential weighting of the graphic element content confidence scores relative to the text content confidence scores, wherein the second differential weighting increases a weighting of the graphic element content confidence score relative to the text content confidence score more than the first differential weighting; and
selecting higher value ones of the first weighted averages and the second weighted averages as the similarity confidence scores for each of the respective ones of the plurality of slides.

3. The method of claim 2 , wherein at least one of the steps of generating the text content confidence scores and generating the graphic element content confidence scores comprises:
generating raw confidence scores as a function of comparing slide content; and
generating content confidence scores by normalizing the raw confidence scores as a function of lowest and highest generated raw score values.

4. The method of claim 2 , wherein the graphic element content confidence scores are generated to represent at least one of:
amounts of similarity of an individual graphic element that is visible within the graphic content of the input slide to individual graphic elements that are visible within the graphic content of the respective ones of the plurality of slides, as a function of at least one of shape, scale, and color attributes; and
amounts of similarity of an arrangement of multiple elements that are each visible within the graphic content of the input slide relative to each other to arrangements of multiple elements that are visible within the graphic content of respective ones of the plurality of slides, as a function of at least one of shape, scale, and color attributes.

5. The method of claim 2 , further comprising:
in response to determining at least one of a lapse of an update period, and a threshold number of identifications of the input slide for comparison of slide text content and graphic element content to the respective ones of the plurality of slides has been met since the similarity confidence scores were most recently generated, repeating the steps of, in response to the identification of an input slide, comparing the slide text content of the input slide to the slide text content of the respective ones of the plurality of slides, comparing the graphic element content of the input slide to the graphic element content of the respective ones of the plurality of slides, generating the text content confidence scores, generating the graphic element content confidence scores, generating the similarity confidence scores, and ranking the plurality of slides for similarity to the input slide as a function of the generated similarity confidence scores.

6. The method of claim 1 , further comprising:
integrating computer-readable program code into a computer system comprising the processor, a computer readable memory in circuit communication with the processor, and a computer readable storage medium in circuit communication with the processor; and
wherein the processor executes program code instructions stored on the computer-readable storage medium via the computer readable memory and thereby performs the steps of generating the text content confidence scores that represent amounts of similarity of slide text content of the input slide to compared text content of the plurality of repository slides in response to the identification of an input slide, generating the graphic element content confidence scores that represent amounts of similarity of graphic content of the input slide to compared graphic element content of the respective ones of the plurality of slides, generating the similarity confidence scores for each of the respective ones of the plurality of slides as functions of weighted averages of the text content confidence scores and the graphic element content confidence scores generated for the respective ones of the plurality of slides, and ranking the plurality of slides for similarity to the input slide as the function of the generated similarity confidence scores.

7. The method of claim 6 , further comprising:
providing the integrated computer-readable program code as a service in a cloud environment.

8. The method of claim 2 , wherein the compared slide text content comprises a plurality of different text contents that comprise slide title text content and slide body text content;
wherein the step of generating the text content confidence scores comprises generating a slide title text content confidence score, and a slide body text content confidence score; and
wherein the generated similarity confidence scores for each of the respective ones of the plurality of slides are functions of weighted averages of the slide title text content confidence scores and the slide body text content confidence scores, wherein the slide title text content confidence scores are weighted differently from the slide body text content confidence scores.

9. The method of claim 8 , wherein the plurality of different text contents comprises slide speakers notes text content; and
wherein the step of generating the text content confidence scores comprises generating a slide speakers notes text content confidence score; and
wherein the generated similarity confidence scores for each of the respective

Assignees

• IBM

Inventors

• Chamberlain James L
• Clinton James H
• Dillard Blain A
• Hoffman David A
• White Terrence E

Classifications

• G06F18/22

  Matching criteria, e.g. proximity measures · CPC title

• G06F40/194 Primary

  Calculation of difference between files · CPC title

• G06F17/2211

  Physics · mapped topic

• G06K9/6215 Primary

  Physics · mapped topic

• G06K9/6212

  Physics · mapped topic