Back to results
US 12,536,227 B2Granted
Information processing apparatus, information processing method, and program
AssigneeSony Group Corporation
InventorWeihsiang Liao
Life of this patent
Estimating the term…
Abstract
An information processing apparatus according to an embodiment includes a search unit that evaluates similarity between two pieces of audio data on the basis of the distance in a latent space between latent codes obtained from the two pieces of audio data.
Claims20
An information processing system comprising:
- circuitry configured to:
- encode each of a plurality of first pieces of audio data registered in a sound library and a second piece of audio data input by a user into respective latent codes;
- reduce dimensions of the respective latent codes to generate respective reduced latent codes in a two-dimensional or three-dimensional latent space;
- evaluate similarity between the second piece of audio data and each of the first pieces of audio data on the basis of a distance in a latent space between the reduced latent code of the second piece of audio data and the reduced latent code of the corresponding first piece of audio data;
- and
- retrieve, from the sound library, one or more of the first pieces of audio data based on the distance.
3 more elements
An information processing method comprising:
- encoding each of a plurality of first pieces of audio data registered in a sound library and a second piece of audio data input by a user into respective latent codes;
- reducing dimensions of the respective latent codes to generate respective reduced latent codes in a two-dimensional or three-dimensional latent space;
- evaluating similarity between the second piece of audio data and each of the first pieces of audio data on the basis of a distance in a latent space between the reduced latent code of the second piece of audio data and the reduced latent code of the corresponding first piece of audio data;
- and
- retrieving, from the sound library, one or more of the first pieces of audio data based on the distance.
2 more elements
A non-transitory computer-readable medium including computer program code, which when executed by circuitry, causes the circuitry to:
- encode each of a plurality of first pieces of audio data registered in a sound library and a second piece of audio data input by a user into respective latent codes;
- reduce dimensions of the respective latent codes to generate respective reduced latent codes in a two-dimensional or three-dimensional latent space;
- evaluate similarity between the second piece of audio data and each of the first pieces of audio data on the basis of a distance in a latent space between the reduced latent code of the second piece of audio data and the reduced latent code of the corresponding first piece of audio data;
- and
- retrieve, from the sound library, one or more of the first pieces of audio data based on the distance.
2 more elements
Dependencies are inferred from the claim text.