US 12,380,893 B1Granted

Enhanced content matching using content fingerprints

AssigneeAmazon Technologies, Inc.

InventorBalaji Nagendran Thoshkahna

Life of this patent

Estimating the term…

Loading patent…

Abstract

Devices, systems, and methods are provided for using fingerprints of signals to match content. A method may include receiving, by a device, a first signal; generating a first value indicative of a first peak of a spectrogram for the first signal; generating a second value indicative of a second peak of the spectrogram; generating a third value indicative of a difference between a first time of the first peak and a second time of the second peak; generating a first quantized value based on the first value; generating a second quantized based on the second value; generating a first fingerprint based on the first quantized value, the second quantized value, and the third value; determining, based on a comparison of the first fingerprint to a second fingerprint, a match between the first fingerprint and the second fingerprint; and sending an indication of the match.

Claims20

Dependent claims

A method, comprising:

receiving, by at least one processor of a device, a first audio signal comprising a portion of a song;
generating, by the at least one processor, a second audio signal based on shifting the first audio signal by a first amount of time in a first direction;
generating, by the at least one processor, a third audio signal based on shifting the first audio signal by a second amount of time in a second direction;
generating, by the at least one processor, a first spectrogram based on the first audio signal;
generating, by the at least one processor, a second spectrogram based on the second audio signal;
generating, by the at least one processor, a third spectrogram based on the third audio signal;
generating, by the at least one processor, an averaged spectrogram based on an average of the first spectrogram, the second spectrogram, and the third spectrogram;
identifying, by the at least one processor, a first peak of the averaged spectrogram, the first peak occurring at a first time;
identifying, by the at least one processor, a second peak of the averaged spectrogram, the second peak occurring at a second time;
generating, by the at least one processor, a first value indicative of the first peak, the first value consisting of a first number of bits;
generating, by the at least one processor, a second value indicative of the second peak, the second value consisting of a second number of bits less than the first number of bits;
generating, by the at least one processor, a third value indicative of a difference between the second time and the first time;
generating, by the at least one processor, a first quantized value by setting a least significant bit from the first value to zero;
generating, by the at least one processor, a second quantized value by setting a least significant bit from the second value to zero;
generating, by the at least one processor, based on a hash using the first quantized value and the second quantized value, and the hash excluding the third value indicative of the difference between the second time and first time, a first fingerprint of the first quantized value, the second quantized value, and the third value;
determining, by the at least one processor, based on a comparison of the first fingerprint to a second fingerprint generated based on a fourth audio signal, a match between the first fingerprint and the second fingerprint;
determining, by the at least one processor, based on the match, that the first fingerprint and the second fingerprint are based on the song;
and
sending, by the at least one processor, an indication that the first fingerprint and the second fingerprint are based on the song.

16 more elements

A method, comprising:

receiving, by at least one processor of a device, a first signal;
generating, by the at least one processor, a second signal by time-shifting the first signal;
generating, by the at least one processor, an averaged spectrogram by averaging a first spectrogram for the first signal and a second spectrogram for the second signal;
generating, by the at least one processor, a first value indicative of a first peak of the averaged spectrogram, the first peak occurring at a first time;
generating, by the at least one processor, a second value indicative of a second peak of the averaged spectrogram, the second peak occurring at a second time;
generating, by the at least one processor, a third value indicative of a difference between the second time and the first time;
generating, by the at least one processor, at least one quantized value by setting at least one bit from at least one of the first value or the second value to zero;
generating, by the at least one processor, based on a hash using the first value, the second value, and the at least one quantized value, and the hash excluding the third value indicative of the difference between the second time and the first time, a first fingerprint of the first value, the second value, and the third value;
determining, by the at least one processor, based on a comparison of the first fingerprint to a second fingerprint generated based on a third signal, a match between the first fingerprint and the second fingerprint;
and
sending, by the at least one processor, an indication of the match.

8 more elements

A system comprising at least one processor coupled to memory, the at least one processor configured to:

receive a first signal;
generate a second signal by time-shifting the first signal;
generate an averaged spectrogram by averaging a first spectrogram for the first signal and a second spectrogram for the second signal;
generate a first value indicative of a first peak of the averaged spectrogram, the first peak occurring at a first time;
generate a second value indicative of a second peak of the averaged spectrogram, the second peak occurring at a second time;
generate a third value indicative of a difference between the second time and the first time;
generate at least one quantized value by setting at least one bit from at least one of the first value or the second value to zero;
generate, based on a hash using the first value, the second value, and the at least one quantized value, and the hash excluding the third value indicative of the difference between the second time and the first time, a first fingerprint of the first value, the second value, and the third value;
determine, based on a comparison of the first fingerprint to a second fingerprint generated based on a third signal, a match between the first fingerprint and the second fingerprint;
and
send an indication of the match.

8 more elements

Dependencies are inferred from the claim text.