Systems and methods for performing semantic analysis of media objects
First Claim
1. A method for restricting access to digital video content based on recognition of a TV station logo embedded within the video content that varies according to the video usage rights, the method comprising:
- extracting the TV station logo from the video content by combining several frames of video content;
identifying the TV station logo by comparing the TV station logo to known TV station logos that most closely matches said TV station logo extracted from the video content, where the known TV station logos are associated with user profiles of copyright rights holders, and wherein at least some of the known TV station logos are supplied by rights holders for content analysis to identify their owned content;
estimating a first match probability of the likely copyright rights holder based on the identified TV station logo;
identifying a second media object within at least one frame of the video content different from the TV station logo;
extracting at least a second semantic property from the second specified media object within at least one frame of the video content;
matching said at least second semantic property of the second specified media object against the known semantic properties of known media objects associated with user profiles of copyright rights holders;
comparing matches between said at least second semantic property and the known semantic properties to a second threshold value;
estimating a second match probability of the likely copyright rights holder based on the second compared matches;
combining the first match probability and the second match probability into a combined match probability;
identifying a copyright rights holder based on at least the combined match probability;
retrieving rules governing distribution rights of the content based at least on identification of the copyright rights holder; and
restricting access to the digital video content in accordance with the retrieved rules.
5 Assignments
0 Petitions
Accused Products
Abstract
System and method for identification of ownership to media objects through analysis of semantic information in the media objects are disclosed Semantic information is a term that can be used to describe aspects of the content of a media object that are apparent to a human viewing or listening to the content One embodiment of the invention includes extracting semantic information from a specified media object, matching said semantic information of the specified media object against known semantic information of known media objects, and taking action based on the result of the companson.
23 Citations
23 Claims
-
1. A method for restricting access to digital video content based on recognition of a TV station logo embedded within the video content that varies according to the video usage rights, the method comprising:
-
extracting the TV station logo from the video content by combining several frames of video content; identifying the TV station logo by comparing the TV station logo to known TV station logos that most closely matches said TV station logo extracted from the video content, where the known TV station logos are associated with user profiles of copyright rights holders, and wherein at least some of the known TV station logos are supplied by rights holders for content analysis to identify their owned content; estimating a first match probability of the likely copyright rights holder based on the identified TV station logo; identifying a second media object within at least one frame of the video content different from the TV station logo; extracting at least a second semantic property from the second specified media object within at least one frame of the video content; matching said at least second semantic property of the second specified media object against the known semantic properties of known media objects associated with user profiles of copyright rights holders; comparing matches between said at least second semantic property and the known semantic properties to a second threshold value; estimating a second match probability of the likely copyright rights holder based on the second compared matches; combining the first match probability and the second match probability into a combined match probability; identifying a copyright rights holder based on at least the combined match probability; retrieving rules governing distribution rights of the content based at least on identification of the copyright rights holder; and restricting access to the digital video content in accordance with the retrieved rules. - View Dependent Claims (2, 3, 4, 15, 16, 19, 21)
-
-
5. A method for restricting access to a media object based on automatic recognition of human recognizable semantic properties evident during playback of the media object, the method comprising:
-
extracting at least one semantic property from a first specified media object within at least one frame of video content; matching said at least one semantic property of the first specified media object against known semantic properties of known media objects associated with user profiles of rights holders, wherein at least some of the known semantic properties are generated from material supplied by rights holders for content analysis to identify their owned content; comparing matches between said at least one semantic property and known semantic properties to a threshold value; estimating a first match probability of the likely copyright rights holder based on the compared matches; identifying a second media object within at least one frame of the video content different from the first media object; extracting at least a second semantic property from the second specified media object within at least one frame of the video content; matching said at least second semantic property of the second specified media object against the known semantic properties of known media objects associated with user profiles of copyright rights holders; comparing matches between said at least second semantic property and the known semantic properties to a second threshold value; estimating a second match probability of the likely copyright rights holder based on the second compared matches; combining the first match probability and the second match probability into a combined match probability; identifying a copyright rights holder based on at least the combined match probability; retrieving rules governing distribution rights of the video content based at least on identification of the copyright rights holder; and determining allowed actions that can be performed with respect to the media objects based upon the retrieved rules; restricting access to playback of video content containing the media objects in accordance with the allowed actions. - View Dependent Claims (6, 7, 8, 9, 10, 11, 12, 13, 14, 17, 18, 20, 22, 23)
-
Specification