Method and apparatus for using confidence scores of enhanced metadata in search-driven media applications
According to one aspect, a computerized method and apparatus for generating and presenting search snippets that enable user-directed navigation of the underlying audio/video content. The method involves obtaining metadata associated with discrete media content that satisfies a search query. The metadata identifies a number of content segments and corresponding timing information derived from the underlying media content using one or more automated media processing techniques. Using the timing information identified in the metadata, a search result or “snippet” can be generated that enables a user to arbitrarily select and commence playback of the underlying media content at any of the individual content segments.
1 - 24 . (canceled)
25 . A computerized method of generating search results for media content, comprising
obtaining a metadata document corresponding to media content from a search query, the metadata document including text recognized from an audio portion of the media content and one or more confidence scores associated with the recognized text; and
determining whether to identify the media content or a portion of the media content in a search result based on the one or more confidence scores from the metadata document.
26 . The computerized method of claim 25 wherein the one or more confidence scores represent the accuracy of the recognized text.
27 . The computerized method of claim 25 wherein the one or more confidence scores includes a plurality of individual confidence scores corresponding to the text for each spoken word recognized from the audio portion of the media content.
28 . The computerized method of claim 25 wherein the one or more confidence scores includes a plurality of confidence scores corresponding to segments of the media content, each of the segment confidence scores being derived from the individual confidence scores of the text comprising the segment.
29 . The method of claim 25 wherein the one or more confidence scores includes an overall confidence score derived from the individual confidence scores of substantially all of the text recognized from the audio portion of the media content.
30 . The method of claim 27 , further comprising:
generating a search result that includes a portion of the recognized text, the text for one or more spoken words having an individual confidence score that fails to satisfy a predefined threshold is omitted or replaced with one or more predefined symbols.
31 . The computerized method of claim 27 , wherein the metadata document groups portions of the recognized text according to content segments, and the method further comprising:
deriving a confidence score for at least one of the content segments from the individual confidence scores of the recognized text that comprise the at least one content segment; and
determining whether to include the at least one content segment in the search result from the confidence score derived for the at least one content segment.
32 . The computerized method of claim 31 further comprising:
excluding the at least one content segment that has a confidence score failing to satisfy a predefined threshold from the search result.
33 . The computerized method of claim 31 wherein one or more of the content segments of the metadata include word segments, audio speech segments, video segments, non-speech audio segments, or marker segments.
34 . The computerized method of claim 27 , further comprising:
deriving an overall confidence score from the individual confidence scores of substantially all of the recognized text from the audio portion of the media content; and
determining whether to identify the media content in a search result from the overall confidence score.
35 . The computerized method of claim 34 further comprising:
excluding the identity of the media content having an overall confidence score failing to satisfy a predefined threshold from the search result.
36 . A computerized method of generating search results for media content, comprising
obtaining a plurality of metadata documents corresponding to a plurality of media content from a search query, each of the plurality of metadata documents including text recognized from an audio portion of corresponding media content and one or more confidence scores associated with the recognized text; and
determining a ranking order of the plurality of media content according to one or more factors, at least one of the factors based on the one or more confidence scores from the plurality of metadata documents.
37 . The computerized method of claim 36 , further comprising:
sorting the plurality of metadata documents according to the determined ranking order;
generating a plurality of search results ordered according to the sorted plurality of metadata documents.
38 . The computerized method of claim 36 , wherein each of the plurality of metadata documents groups portions of the recognized text according to content segments, and the method further comprising:
for each of the plurality of metadata documents, deriving a confidence score from at least one of the content segments that is derived from the individual confidence scores of the recognized text that comprise the at least one content segment; and
determining a ranking order of the plurality of media content according to one or more factors, at least one of the factors including the confidence score from at least one of the content segments of the plurality of metadata documents.
39 . The computerized method of claim 38 wherein one or more of the content segments identified in the metadata document include word segments, audio speech segments, video segments, non-speech audio segments, or marker segments.
40 . A computerized apparatus for generating search results for media content, comprising
means for obtaining a metadata document corresponding to media content from a search query, the metadata document including text recognized from an audio portion of the media content and one or more confidence scores associated with the recognized text; and
means for determining whether to identify the media content or a portion of the media content in a search result based on the one or more confidence scores from the metadata document.
41 . The computerized apparatus of claim 40 wherein the one or more confidence scores includes a plurality of individual confidence scores corresponding to the text for each spoken word recognized from the audio portion of the media content.
42 . The computerized apparatus of claim 40 wherein the one or more confidence scores includes a plurality of confidence scores corresponding to segments of the media content, each of the segment confidence scores being derived from the individual confidence scores of the text comprising the segment.
43 . The computerized apparatus of claim 40 wherein the one or more confidence scores includes an overall confidence score derived from the individual confidence scores of substantially all of the text recognized from the audio portion of the media content.
44 . The computerized apparatus of claim 41 , further comprising:
generating a search result that includes a portion of the recognized text, the text for one or more spoken words having an individual confidence score that fails to satisfy a predefined threshold is omitted or replaced with one or more predefined symbols.
45 . The computerized apparatus of claim 41 , wherein the metadata document groups portions of the recognized text according to content segments, and the method further comprising:
means for deriving a confidence score for at least one of the content segments from the individual confidence scores of the recognized text that comprise the at least one content segment; and
means for determining whether to include the at least one content segment in the search result from the confidence score derived for the at least one content segment.
46 . The computerized apparatus of claim 41 , further comprising:
means for deriving an overall confidence score from the individual confidence scores of substantially all of the recognized text from the audio portion of the media content; and
means for determining whether to identify the media content in a search result from the overall confidence score.
47 . A computerized apparatus of generating search results for media content, comprising
means for obtaining a plurality of metadata documents corresponding to a plurality of media content from a search query, each of the plurality of metadata documents including text recognized from an audio portion of corresponding media content and one or more confidence scores associated with the recognized text; and
means for determining a ranking order of the plurality of media content according to one or more factors, at least one of the factors based on the one or more confidence scores from the plurality of metadata documents.
48 . The computerized apparatus of claim 47 , further comprising:
means for sorting the plurality of metadata documents according to the determined ranking order;
means for generating a plurality of search results ordered according to the sorted plurality of metadata documents.
49 . The computerized apparatus of claim 47 , wherein each of the plurality of metadata documents groups portions of the recognized text according to content segments, and the method further comprising:
for each of the plurality of metadata documents, means for deriving a confidence score from at least one of the content segments that is derived from the individual confidence scores of the recognized text that comprise the at least one content segment; and
means for determining a ranking order of the plurality of media content according to one or more factors, at least one of the factors including the confidence score from at least one of the content segments of the plurality of metadata documents.
50 . The computerized apparatus of claim 49 wherein one or more of the content segments identified in the plurality of metadata documents include word segments, audio speech segments, video segments, non-speech audio segments, or marker segments.