Baftiu, Naim and Bytyci, Dionesa and Atanasova-Pacemska, Tatjana and Atanasova, Ana and Lameski, Petre and Zdravevski, Eftim (2026) Robust speech enhancement in noisy environments using deep convolutional autoencoders supported by HPC. Prezglad Elektrotechniczny, 102 (8/2026). pp. 51-55. ISSN 0033-2097 / 2449-9544 (online)
![[thumbnail of Trud Polska, Scopus 8.2026..pdf]](https://eprints.ugd.edu.mk/style/images/fileicons/text.png "Trud Polska, Scopus 8.2026..pdf")
Text
Trud Polska, Scopus 8.2026..pdf
Download (1MB)
Official URL: https://pe.org.pl/en/
Abstract
This paper presents a hardware-accelerated Deep Convolutional Autoencoder for speech enhancement. Using a U-Net architecture on spectrograms,
the proposed model achieves a 79.08% reduction in Mean Squared Error (MSE). The study demonstrates that leveraging High-Performance
Computing (HPC) significantly improves training efficiency and signal fidelity in noisy environments.
| Item Type: | Article |
|---|---|
| Subjects: | Natural sciences > Computer and information sciences Engineering and Technology > Electrical engineering, electronic engineering, information engineering Natural sciences > Matematics |
| Divisions: | Faculty of Computer Science |
| Depositing User: | Tatjana A. Pacemska |
| Date Deposited: | 24 Sep 2026 08:04 |
| Last Modified: | 24 Sep 2026 08:04 |
| URI: | https://eprints.ugd.edu.mk/id/eprint/39109 |
Actions (login required)
- View Item