Media Fingerprints that Reliably Correspond to Media Content with Projection of Moment Invariants
AssigneeDolby Laboratories Licensing Corporation
InventorsRegunathan Radhakrishnan, Claus Bauer
Life of this patent
Estimating the term…
Abstract
A portion of media content is accessed. Components from a first and each subsequent spatial regions of the media content are sampled. Each spatial region has an unsegmented area. Each subsequent spatial region includes those within its area as elements thereof or the spatial regions may partially overlap. The regions may overlap independent of a hierarchical relationship between the regions. A media fingerprint is derived from the components of each of the spatial regions, which reliably corresponds to the media content portion, e.g., over geometric attacks such as rotation.
Claims32
A method, comprising the steps of:
- accessing a portion of media content;
- wherein the media content comprises one or more of a video image or an audio spectrogram;
- sampling one or more components from a first spatial region of the media content;
- sampling one or more components from a second spatial region of the media content;
- wherein the first spatial region overlaps with the second spatial region;
- deriving a first media fingerprint from the components of the first and the second spatial regions based on moment invariants and/or based on chrominance related and luminance related values;
- generating a difference instance from one or more of the video image or the audio spectrogram by comparing a current version and a temporally averaged version thereof;
- computing a plurality of moment invariants from a set of regions in the difference instance;
- deriving a second media fingerprint based on the plurality of computed moment invariants;
- and
- concatenating the first media fingerprint and the second media fingerprint.
Dependencies are inferred from the claim text.