Research Grant in Silent Speech Recognition and Biomedical Signal Processing

This research grant at Universidade da Madeira focuses on advancing assistive communication technologies for voiceless individuals. The successful candidate will investigate silent speech recognition using surface electromyography (sEMG) and develop AI models to translate facial and articulatory muscle activity into comprehensible language. The project aims to create wearable, real-time devices that enable communication for those with speech impairments. The work will be conducted at the Faculty of Exact Sciences and Engineering under the supervision of Sheikh Shanawaz Mostafa and Morgado Dias, with a strong emphasis on interdisciplinary collaboration between biomedical engineering, computer engineering, and artificial intelligence.

• Conduct a literature review on silent speech recognition, sEMG, and biomedical signal processing.
• Collect, organize, and preprocess facial and articulatory sEMG signal databases during silent speech tasks.
• Develop preprocessing pipelines for noise removal, signal segmentation, normalization, and feature extraction.
• Design, train, and evaluate machine learning and deep learning models (e.g., SVM, CNN, RNN, Transformer-based models) for word and phrase recognition.
• Implement and test language modeling methods to improve phrase prediction and reduce recognition errors.
• Validate the proposed system experimentally using metrics such as accuracy, robustness, user comfort, and real-time performance.
• Study the adaptability of the system to different users with minimal calibration.
• Write scientific articles, disseminate research results, and prepare the doctoral thesis.