Speech Enhancement for Hearing Impaired Based on Bandpass Filters and a Compound Deep Denoising Autoencoder
Deep neural networks have been applied for speech enhancements efficiently. However, for large variations of speech patterns and noisy environments, an individual neural network with a fixed number of hidden layers causes strong interference, which can lead to a slow learning process, poor generalis...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-07-01
|
| Series: | Symmetry |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2073-8994/13/8/1310 |
| _version_ | 1797522026547642368 |
|---|---|
| author | Raghad Yaseen Lazim AL-Taai Xiaojun Wu |
| author_facet | Raghad Yaseen Lazim AL-Taai Xiaojun Wu |
| author_sort | Raghad Yaseen Lazim AL-Taai |
| collection | DOAJ |
| description | Deep neural networks have been applied for speech enhancements efficiently. However, for large variations of speech patterns and noisy environments, an individual neural network with a fixed number of hidden layers causes strong interference, which can lead to a slow learning process, poor generalisation in an unknown signal-to-noise ratio in new inputs, and some residual noise in the enhanced output. In this paper, we present a new approach for the hearing impaired based on combining two stages: (1) a set of bandpass filters that split up the signal into eight separate bands each performing a frequency analysis of the speech signal; (2) multiple deep denoising autoencoder networks, with each working for a small specific enhancement task and learning to handle a subset of the whole training set. To evaluate the performance of the approach, the hearing-aid speech perception index, the hearing aid sound quality index, and the perceptual evaluation of speech quality were used. Improvements in speech quality and intelligibility were evaluated using seven subjects of sensorineural hearing loss audiogram. We compared the performance of the proposed approach with individual denoising autoencoder networks with three and five hidden layers. The experimental results showed that the proposed approach yielded higher quality and was more intelligible compared with three and five layers. |
| first_indexed | 2024-03-10T08:20:48Z |
| format | Article |
| id | doaj.art-ba6bb2c51428419a9d414f1ae359b8c9 |
| institution | Directory Open Access Journal |
| issn | 2073-8994 |
| language | English |
| last_indexed | 2024-03-10T08:20:48Z |
| publishDate | 2021-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Symmetry |
| spelling | doaj.art-ba6bb2c51428419a9d414f1ae359b8c92023-11-22T09:59:13ZengMDPI AGSymmetry2073-89942021-07-01138131010.3390/sym13081310Speech Enhancement for Hearing Impaired Based on Bandpass Filters and a Compound Deep Denoising AutoencoderRaghad Yaseen Lazim AL-Taai0Xiaojun Wu1School of Computer Science, Shaanxi Normal University, Xi’an 710119, ChinaSchool of Computer Science, Shaanxi Normal University, Xi’an 710119, ChinaDeep neural networks have been applied for speech enhancements efficiently. However, for large variations of speech patterns and noisy environments, an individual neural network with a fixed number of hidden layers causes strong interference, which can lead to a slow learning process, poor generalisation in an unknown signal-to-noise ratio in new inputs, and some residual noise in the enhanced output. In this paper, we present a new approach for the hearing impaired based on combining two stages: (1) a set of bandpass filters that split up the signal into eight separate bands each performing a frequency analysis of the speech signal; (2) multiple deep denoising autoencoder networks, with each working for a small specific enhancement task and learning to handle a subset of the whole training set. To evaluate the performance of the approach, the hearing-aid speech perception index, the hearing aid sound quality index, and the perceptual evaluation of speech quality were used. Improvements in speech quality and intelligibility were evaluated using seven subjects of sensorineural hearing loss audiogram. We compared the performance of the proposed approach with individual denoising autoencoder networks with three and five hidden layers. The experimental results showed that the proposed approach yielded higher quality and was more intelligible compared with three and five layers.https://www.mdpi.com/2073-8994/13/8/1310compound neural networkdeep denoising autoencoderhearing aid applicationbandpass filterdeep learning |
| spellingShingle | Raghad Yaseen Lazim AL-Taai Xiaojun Wu Speech Enhancement for Hearing Impaired Based on Bandpass Filters and a Compound Deep Denoising Autoencoder Symmetry compound neural network deep denoising autoencoder hearing aid application bandpass filter deep learning |
| title | Speech Enhancement for Hearing Impaired Based on Bandpass Filters and a Compound Deep Denoising Autoencoder |
| title_full | Speech Enhancement for Hearing Impaired Based on Bandpass Filters and a Compound Deep Denoising Autoencoder |
| title_fullStr | Speech Enhancement for Hearing Impaired Based on Bandpass Filters and a Compound Deep Denoising Autoencoder |
| title_full_unstemmed | Speech Enhancement for Hearing Impaired Based on Bandpass Filters and a Compound Deep Denoising Autoencoder |
| title_short | Speech Enhancement for Hearing Impaired Based on Bandpass Filters and a Compound Deep Denoising Autoencoder |
| title_sort | speech enhancement for hearing impaired based on bandpass filters and a compound deep denoising autoencoder |
| topic | compound neural network deep denoising autoencoder hearing aid application bandpass filter deep learning |
| url | https://www.mdpi.com/2073-8994/13/8/1310 |
| work_keys_str_mv | AT raghadyaseenlazimaltaai speechenhancementforhearingimpairedbasedonbandpassfiltersandacompounddeepdenoisingautoencoder AT xiaojunwu speechenhancementforhearingimpairedbasedonbandpassfiltersandacompounddeepdenoisingautoencoder |