Back to results
US 2022/0092434 A1Published application
Noise waveform removing device, model training device, noise waveform removing method, model training method, generation model, and wearable device
AssigneeCasio Computer Co., Ltd.
InventorMasashi UEDA
Life of this patent
Estimating the term…
Abstract
A noise waveform removing device includes one or more processors that are configured to obtain noise-containing movement waveform data; use a trained generation model, to determine a filter contribution rate based on the noise-containing movement waveform data; and generate noise-free movement waveform data, based on the determined filter contribution rate and the noise-containing movement waveform data.
Claims18
A noise waveform removing device comprising:
- one or more processors,
- wherein the one or more processors are configured to
- obtain noise-containing movement waveform data;
- use a trained generation model, to determine a filter contribution rate based on the noise-containing movement waveform data;
- and
- generate noise-free movement waveform data, based on the determined filter contribution rate and the noise-containing movement waveform data.
3 more elements
A model training device comprising:
- one or more processors;
- and
- one or more memories,
- wherein the one or more memories are configured to store training data including noise-containing movement waveform data for training and noise-free movement waveform data for training, and
- wherein the one or more processors are configured to train a generation model that determines a filter contribution rate based on the noise-containing movement waveform data for training, and a discrimination model that discriminates, from the training data, pseudo noise-free movement waveform data generated based on the filter contribution rate and the noise-containing movement waveform data for training, according to a GAN (Generative Adversarial Network).
2 more elements
A noise waveform removing method comprising:
- obtaining noise-containing movement waveform data, executed by one or more processors;
- using a trained generation model, to determine a filter contribution rate based on the noise-containing movement waveform data, executed by the one or more processors;
- and
- generating noise-free movement waveform data, based on the determined filter contribution rate and the noise-containing movement waveform data, executed by the one or more processors.
1 more element
A model training method comprising:
- storing training data including noise-containing movement waveform data for training and noise-free movement waveform data for training, executed by one or more processors;
- and
- training a generation model that determines a filter contribution rate based on the noise-containing movement waveform data for training, and a discrimination model that discriminates, from the training data, pseudo noise-free movement waveform data generated based on the filter contribution rate and the noise-containing movement waveform data for training, according to a GAN (Generative Adversarial Network), executed by the one or more processors.
Dependencies are inferred from the claim text.