Convolutional Variational Autoencoders for Spectrogram Compression in Automatic Speech Recognition
Fuente:
arXiv
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912058258227200 |
|---|---|
| author | Iakovenko, Olga Bondarenko, Ivan |
| author_facet | Iakovenko, Olga Bondarenko, Ivan |
| contents | For many Automatic Speech Recognition (ASR) tasks audio features as spectrograms show better results than Mel-frequency Cepstral Coefficients (MFCC), but in practice they are hard to use due to a complex dimensionality of a feature space. The following paper presents an alternative approach towards generating compressed spectrogram representation, based on Convolutional Variational Autoencoders (VAE). A Convolutional VAE model was trained on a subsample of the LibriSpeech dataset to reconstruct short fragments of audio spectrograms (25 ms) from a 13-dimensional embedding. The trained model for a 40-dimensional (300 ms) embedding was used to generate features for corpus of spoken commands on the GoogleSpeechCommands dataset. Using the generated features an ASR system was built and compared to the model with MFCC features. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_02560 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Convolutional Variational Autoencoders for Spectrogram Compression in Automatic Speech Recognition Iakovenko, Olga Bondarenko, Ivan Sound Computation and Language Audio and Speech Processing For many Automatic Speech Recognition (ASR) tasks audio features as spectrograms show better results than Mel-frequency Cepstral Coefficients (MFCC), but in practice they are hard to use due to a complex dimensionality of a feature space. The following paper presents an alternative approach towards generating compressed spectrogram representation, based on Convolutional Variational Autoencoders (VAE). A Convolutional VAE model was trained on a subsample of the LibriSpeech dataset to reconstruct short fragments of audio spectrograms (25 ms) from a 13-dimensional embedding. The trained model for a 40-dimensional (300 ms) embedding was used to generate features for corpus of spoken commands on the GoogleSpeechCommands dataset. Using the generated features an ASR system was built and compared to the model with MFCC features. |
| title | Convolutional Variational Autoencoders for Spectrogram Compression in Automatic Speech Recognition |
| topic | Sound Computation and Language Audio and Speech Processing |
| url | https://arxiv.org/abs/2410.02560 |