Segmentation Of Video According To Narrative Theme
Abstract
Content items may be segmented and labeled by a theme to provide information, such as audio or video, with greater functionality, accuracy and speed. The segments and other related information may be stored and made accessible to users through, for example, a search service and/or an on-demand service. Also provided is automatic segmentation that may include any one of receiving a textual description of a content item, determining themes within the textual description, determining relevance intervals corresponding to terms of the theme, filtering and/or merging the relevance intervals, and determining a total interval for each theme that represents one or more contextually-coherent segments of the content item.
Claims
exact text as granted — not AI-modified1 . A method comprising:
determining, by a computing device and for each term selected from a dialog of a content item, a relevance interval for the term by identifying one or more portions of the dialog that are relevant to the term based on a similarity measurement for a first term in the one or more portions, the first term being different from the term, wherein the determining results in relevance intervals for terms selected from the dialog; determining, from a textual description that indicates one or more narrative themes of the content item, a first plurality of terms that indicate a first narrative theme of the one or more narrative themes; determining relevance intervals for the first plurality of terms by searching, based on the first plurality of terms, the relevance intervals for the terms selected from the dialog; determining, based on the relevance intervals for the first plurality of terms, a first composite relevance interval that indicates one or more first portions, of the content item, that are relevant to the first narrative theme; receiving, after storing the first composite relevance interval as part of a searchable index, a query indicative of criteria for searching the searchable index; determining, based on the criteria and the searchable index, the first composite relevance interval; and causing, based on the query, display of at least one of the one or more first portions.
2 . The method of claim 1 , wherein the similarity measurement comprises a statistical measurement of term co-occurrence in corpus data, wherein a transcript of the content item comprises the dialog, and wherein the method further comprises:
receiving the transcript of the content item; and storing the relevance intervals for the terms selected form the dialog in a database; wherein the textual description comprises a user-provided description of the content item; wherein the searching, based on the terms indicative of the first narrative theme, the relevance intervals for the terms selected from the dialog comprises determining that one of the terms selected from the dialog matches one of the first plurality of terms.
3 . The method of claim 1 , further comprising:
determining, from the textual description, a second plurality of terms that indicate a second narrative theme of the one or more narrative themes; determining relevance intervals for the second plurality of terms by searching, based on the second plurality of terms, the relevance intervals for the terms selected from the dialog; determining a second composite relevance interval by removing a fourth relevance interval from the relevance intervals for the second plurality of terms or by merging a fifth relevance interval from the relevance intervals for the second plurality of terms with a sixth relevance interval from the relevance intervals for the second plurality of terms; and after determining the first composite relevance interval, filtering the first composite relevance interval by determining that a first sub-interval of the first composite relevance interval overlaps with a second sub-interval of the second composite relevance interval and by removing the first sub-interval from the first composite relevance interval.
4 . The method of claim 3 , further comprising:
after determining the relevance intervals for the first plurality of terms, filtering the relevance intervals for the first plurality of terms; after determining the relevance intervals for the second plurality of terms, filtering the relevance intervals for the second plurality of terms; after determining the relevance intervals for the first plurality of terms, merging the relevance intervals for the first plurality of terms; after determining the relevance intervals for the second plurality of terms, merging the relevance intervals for the second plurality of terms; after determining the second composite relevance interval, filtering the second composite relevance interval; after determining the first composite relevance interval, merging the first composite relevance interval; and after determining the second composite relevance interval, merging the second composite relevance interval.
5 . The method of claim 1 , wherein determining the first composite relevance interval is performed by removing a first relevance interval from the relevance intervals for the first plurality of terms or by merging a second relevance interval from the first plurality of terms with a third relevance interval from the relevance intervals for the first plurality of terms, and wherein the method further comprises:
preparing the searchable index to associate the first composite relevance interval with the first narrative theme; and wherein the query indicates the first narrative theme based on the first narrative theme being associated with content currently being viewed by a user.
6 . The method of claim 5 , wherein determining the first composite relevance interval is performed by:
removing the first relevance interval from the relevance intervals for the first plurality of terms based on a size of the first relevance interval, an overlap of the first relevance interval with another relevance interval of the relevance intervals for the first plurality of terms, an adjacency of the first relevance interval with another relevance interval of the relevance intervals for the first plurality of terms, or a content item location of the first relevance interval.
7 . The method of claim 5 , wherein determining the first composite relevance interval is performed by:
merging the second relevance interval with the third relevance interval based on an overlap of the second relevance interval with the third relevance interval or an adjacency of the second relevance interval with the third relevance interval.
8 . The method of claim 1 , wherein each of the one or more narrative themes indicates one of a plurality of plot lines for the content item, wherein the first narrative theme indicates a first plot line of the plurality of plot lines, and wherein the criteria indicates the first plot line.
9 . The method of claim 1 , wherein each of the one or more narrative themes indicates one of a plurality of news segments for the content item, wherein the first narrative theme indicates a first news segment of the plurality of news segments, and wherein the criteria indicates the first news segment.
10 . The method of claim 1 , wherein the criteria indicates a content consumption habit of a user.
11 . A method comprising:
determining, by a computing device and for each term selected from a dialog of a content item, a relevance interval for the term by identifying one or more portions of the dialog that are relevant to the term based on a similarity measurement for a first term in the one or more portions, the first term being different from the term, wherein the determining results in relevance intervals for terms selected from the dialog; receiving a summary of plot lines for the content item; determining, from the summary of the plot lines, a first plurality of terms that indicate a first plot line of the plot lines; determining relevance intervals for the first plurality of terms by searching, based on the first plurality of terms, the relevance intervals for the terms selected from the dialog; determining, based on the relevance intervals for the first plurality of terms, a first composite relevance interval that indicates one or more first portions, of the content item, that are relevant to the first plot line; receiving, after storing the first composite relevance interval as part of a searchable index, a query indicative of criteria for searching the searchable index; determining, based on the criteria and the searchable index, the first composite relevance interval; and causing, based on the query, display of at least one of the one or more first portions.
12 . The method of claim 11 , wherein the receiving the summary of the plot lines is performed based on receiving electronic program guide (EPG) data associated with the content item, wherein the EPG data comprises the summary of the plot lines.
13 . The method of claim 11 , wherein the receiving the summary of the plot lines is performed based on accessing a website that indicates the summary of the plot lines.
14 . The method of claim 11 , wherein the receiving the summary of the plot lines is performed based on downloading the summary of the plot lines from a web server.
15 . The method of claim 11 , wherein the criteria indicates a content consumption habit of a user.
16 . The method of claim 11 , wherein determining the first composite relevance interval is performed by removing a first relevance interval from the relevance intervals for the first plurality of terms or by merging a second relevance interval from the first plurality of terms with a third relevance interval from the relevance intervals for the first plurality of terms, and wherein the method further comprises:
preparing the searchable index to associate the first composite relevance interval with the first plot line; and wherein the criteria is based on content currently being viewed by a user.
17 . The method of claim 11 , further comprising:
determining, from the summary of the plot lines, a second plurality of terms that indicate a second plot line of the plot lines; determining relevance intervals for the second plurality of terms by searching, based on the second plurality of terms, the relevance intervals for the terms selected from the dialog; determining a second composite relevance interval by removing a fourth relevance interval from the relevance intervals for the second plurality of terms or by merging a fifth relevance interval from the relevance intervals for the second plurality of terms with a sixth relevance interval from the relevance intervals for the second plurality of terms; and after determining the first composite relevance interval, filtering the first composite relevance interval by determining that a first sub-interval of the first composite relevance interval overlaps with a second sub-interval of the second composite relevance interval and by removing the first sub-interval from the first composite relevance interval.
18 . A method comprising:
determining, by a computing device and for each term selected from a dialog of a content item, a relevance interval for the term by identifying one or more portions of the dialog that are relevant to the term based on a similarity measurement for a first term in the one or more portions, the first term being different from the term, wherein the determining results in relevance intervals for terms selected from the dialog; receiving a textual summary of the content item; determining, from the textual summary, a first plurality of terms that indicate a first theme of the content item; determining relevance intervals for the first plurality of terms by searching, based on the first plurality of terms, the relevance intervals for the terms selected from the dialog; determining, based on the relevance intervals for the first plurality of terms, a first composite relevance interval that indicates one or more first portions, of the content item, that are relevant to the first theme; receiving, after storing the first composite relevance interval as part of a searchable index, a query indicative of criteria for searching the searchable index; determining, based on the criteria and the searchable index, the first composite relevance interval; and causing, based on the query, display of at least one of the one or more first portions.
19 . The method of claim 18 , wherein the receiving the textual summary of the content item is performed based on receiving electronic program guide (EPG) data associated with the content item, wherein the EPG data comprises the textual summary.
20 . The method of claim 18 , further comprising:
determining, from the textual summary, a second plurality of terms that indicate a second theme of the content item; determining relevance intervals for the second plurality of terms by searching, based on the second plurality of terms, the relevance intervals for the terms selected from the dialog; determining a second composite relevance interval by removing a fourth relevance interval from the relevance intervals for the second plurality of terms or by merging a fifth relevance interval from the relevance intervals for the second plurality of terms with a sixth relevance interval from the relevance intervals for the second plurality of terms; and after determining the first composite relevance interval, filtering the first composite relevance interval by determining that a first sub-interval of the first composite relevance interval overlaps with a second sub-interval of the second composite relevance interval and by removing the first sub-interval from the first composite relevance interval.Join the waitlist — get patent alerts
Track US2020192935A1 — get alerts on status changes and closely related new filings.
We store only your email — no account needed. See our privacy policy.