Noise based interest point density pruning
AssigneeGoogle Inc.
InventorsGeorge Tzanetakis, Dominik Roblek, Matthew Sharifi
Life of this patent
Estimating the term…
Abstract
Systems and methods for noise based interest point density pruning are disclosed herein. The systems include determining an amount of noise in an audio sample and adjusting the amount of interest points within an audio sample fingerprint based on the amount of noise. Samples containing high amounts of noise correspondingly generate fingerprints with more interest points. The disclosed systems and methods allow reference fingerprints to be reduced in size while increasing the size of sample fingerprints. The benefits in scalability do not compromise the accuracy of an audio matching system using noise based interest point density pruning.
Claims16
A system comprising:
- a processor;
- and
- a memory communicatively coupled to the processor, the memory having stored therein computer executable components comprising:
- a reference sample component that determines an amount of interest points in a set of reference fingerprints of a reference sample;
- a noise detection component that determines a noise level in an audio sample;
- an interest point detection component that generates a set of interest points for the audio sample;
- a density component that selects a subset of the interest points, wherein an amount of the interest points to include in the subset is determined as a function of the noise level and the amount of interests points in the set of fingerprints of the reference sample, the amount of interest points to include in the subset increases as the noise level increases;
- and
- a fingerprint component that generates a set of sample fingerprints for the audio sample based on the subset of the interest points.
A method, comprising:
- determining, by a system including a processor, an amount of interest points in a set of reference fingerprints of a reference sample;
- determining, by the system, a noise level in an audio sample;
- generating, by the system, a set of interest points for the audio sample;
- selecting, by the system, a subset of interest points comprising determining an amount of the interest points to include in the subset as a function of the noise level and the amount of interests points in the set of fingerprints of the reference sample, the amount of interest points to include in the subset increases as the noise level increases;
- generating, by the system, a set of sample fingerprints for the audio sample based upon the subset of the interest points.
A non-transitory computer-readable medium having instructions stored thereon that, in response to execution, cause a system including a processor to perform operations comprising:
- determining an amount of interest points in a set of reference fingerprints of a reference sample;
- determining a noise level in an audio sample;
- generating a set of interest points for the audio sample;
- selecting a subset of interest points comprising determining an amount of the interest points to include in the subset as a function of the noise level and the amount of interests points in the set of fingerprints of the reference sample, the amount of interest points to include in the subset increases as the noise level increases;
- generating a set of sample fingerprints for the audio sample based upon the subset of the interest points.
Dependencies are inferred from the claim text.