Systems and methods for correcting errors in caption text
AssigneeRovi Guides, Inc.
InventorsAjay Kumar Gupta, Abhijit Satchidanand Savarkar
Life of this patent
Estimating the term…
Abstract
Systems and methods are described to address shortcomings in conventional systems by correcting an erroneous term in on-screen caption text for a media asset. In some aspects, the systems and methods identify the erroneous term in a text segment of the on-screen caption text, and identify one or more video frames of the media asset corresponding to the text segment. The systems and methods further identify a contextual term related to the erroneous term from the one or more video frames. By accessing a knowledge graph, the systems and methods identify a candidate correction based on the contextual term and a portion of the text segment. Lastly, the systems and methods replaces the erroneous term with the candidate correction.
Claims18
A method for correcting an erroneous term in on-screen caption text of a media asset, comprising:
- analyzing an audio stream of the media asset to determine a first text segment of the on-screen caption text;
- identifying an erroneous term in the first text segment of the on-screen caption text;
- extracting one or more video frames from a video stream of the media asset corresponding to the first text segment;
- analyzing a first video frame of the one or more video frames to determine a contextual term associated with the erroneous term;
- accessing a knowledge graph to identify a candidate correction for the erroneous term based on the contextual term and a portion of the first text segment;
- replacing the erroneous term in the first text segment of the closed captioning text with the candidate correction;
- identifying the erroneous term in a second text segment of the on-screen caption text;
- analyzing a second video frame corresponding to the second text segment to determine a second contextual term associated with the erroneous term;
- accessing the knowledge graph to identify an updated candidate correction based on the first contextual term, the second contextual term, the portion of the first text segment and a portion of the second text segment;
- and
- replacing the erroneous term in the second text segment of the on-screen caption text with the updated candidate correction.
A system for correcting an erroneous term in on-screen caption text of a media asset, comprising:
- a memory storing a knowledge graph;
- and
- control circuitry configured to:
- analyze an audio stream of the media asset to determine a first text segment of the on-screen caption text;
- identify an erroneous term in the first text segment of the on-screen caption text;
- extract one or more video frames from a video stream of the media asset corresponding to the first text segment;
- analyze a first video frame of the one or more video frames to determine a contextual term associated with the erroneous term;
- access a knowledge graph to identify a candidate correction for the erroneous term based on the contextual term and a portion of the first text segment;
- replace the erroneous term in the first text segment of the closed captioning text with the candidate correction;
- identify the erroneous term in a second text segment of the on-screen caption text;
- analyze a second video frame corresponding to the second text segment to determine a second contextual term associated with the erroneous term;
- access the knowledge graph to identify an updated candidate correction based on the first contextual term, the second contextual term, the portion of the first text segment and a portion of the second text segment;
- and
- replace the erroneous term in the second text segment of the on-screen caption text with the updated candidate correction.
Dependencies are inferred from the claim text.