Research projects in biomedical signal processing, multimodal learning, rehabilitation, and computational neuroscience.
MASTER’S THESIS · 2024–2026
Multimodal EEG–sEMG Fusion for Hand Gesture Recognition
Investigated the integration of brain and muscle signals for hand movement recognition in rehabilitation applications. Developed a multimodal learning framework using convolutional neural networks and attention mechanisms to capture complementary information from EEG and sEMG signals. Evaluated its robustness under simulated muscle-fatigue conditions.
Read methods
Methods: EEG and sEMG acquisition and preprocessing, ICA-based artifact reduction, synchronization, wavelet-coherence channel selection, convolutional feature encoders, self-attention, and cross-attention. Fatigue effects were simulated rather than measured clinically.
Explore the EEG–sEMG fusion architecture
Select each stage to explore the method.
↓
↓
↓
Select a stage for a concise explanation.
RESEARCH STUDY · 2026
Fatigue-Resistant sEMG Hand Gesture Recognition
Investigated attention-based and recurrent neural networks for recognizing hand gestures from surface electromyography signals. Evaluated different architectures under simulated fatigue-related signal changes, focusing on recognition robustness for prosthetic control and wearable rehabilitation systems.
Read methods
Methods: 250 ms sEMG windows, conventional classifiers, CNN, GRU, and multi-head self-attention. Analyzed the same recording sessions as the multimodal study, under simulated fatigue-related distortions.
Explore the compared sEMG models
CNN and GRU are compared architectures, not sequential layers.
↓
↓
↓
↓
Select a stage for a concise explanation.
UNDERGRADUATE RESEARCH · SUMMER 2023
Microsaccade Representation in the Visual Nervous System
Studied microsaccades in rodents using gaze recordings. Focused on detecting involuntary eye movements and examining their relationship with visual image statistics.
Read methods
Methods: detecting microsaccadic eye movements in rodent gaze recordings and examining associations with visual image statistics.