Browsing and retrieval of full broadcast-quality video

摘要

A method includes steps of indexing a media collection, searching an indexed library and browsing a set of candidate program segments. The step of indexing a media collection creates the indexed library based on a content of the media collection. The step of searching the indexed library identifies the set of candidate program segments based on a search criteria. The step of browsing the set of candidate program segments selects a segment for viewing.

主权项

1. A method comprising:
classifying a media collection as program content and commercials, to yield classified program content, wherein the commercials are to be presented chronologically separate from the program content and in a same display screen as the program content; recognizing speech in the program content, to yield recognized speech comprising recognized speech timestamps; synchronizing the recognized speech with closed captioning data by aligning the recognized speech timestamps and closed captioning timestamps associated with the closed captioning data, to yield synchronized recognized speech: identifying program segments within the program content based on the synchronized recognized speech in the program content; indexing the media collection based on the program segments to create an indexed library, the indexed library comprising an inverted index of the closed captioning data; receiving a search query to the indexed library from a user; modifying a semantic representation format of the search query to meet a requirement of the indexed library, to yield a modified search query; searching the indexed library and a relational database to identify a set of candidate program segments based on the modified search query, wherein the relational database comprises:
entities comprising a network, a show title, a broadcast date, and a broadcast time of the program content; anda data model specific to the entities within the relational database, wherein the data model comprises attributes belonging to each entity in the entities and relationships between the entities; and transmitting a low-resolution output of the set of candidate program segments for the user to browse and select for high-resolution playback via a user interface.