Resources

Talks

Ongoing Research projects

Emotion recognition using semi-supervised learning from physiological signals

Description: Emotion recognition from physiological signals offers a promising approach for objective assessment of human affective states. My research focuses on developing semi-supervised learning frameworks for emotion recognition using electroencephalogram (EEG) and electrocardiogram (ECG) signals. Since obtaining reliable emotion labels is often expensive, subjective, and time-consuming, the research leverages both labeled and large amounts of unlabeled physiological data to build robust and scalable emotion recognition systems.

The work integrates biomedical signal processing, representation learning, and deep learning techniques to extract discriminative neural and cardiovascular biomarkers associated with emotional states. By employing semi-supervised and domain adaptation approaches, the research aims to improve emotion classification performance while reducing dependence on extensive manual annotations. Applications include affective computing, mental health monitoring, adaptive human-computer interaction, personalized healthcare, and wearable emotion-aware systems.

Epileptic seizure detection using EEG 

Description: This research explores the use of multi-channel Electroencephalography (EEG) signals for the automated detection of epileptic seizures. The focus is on developing intelligent algorithms that analyze complex brain activity patterns to identify the onset of seizure events with high accuracy. By integrating advanced signal processing techniques with machine learning and deep learning models, the system extracts meaningful temporal and spectral features from EEG recordings and distinguishes seizure-related activity from normal brain states. The research aims to support clinicians with reliable, real-time diagnostic tools, reduce the burden of manual EEG analysis, and contribute to the development of next-generation neurological monitoring systems for improved epilepsy management and patient care.

Sleep disorders diagnosis and sleep staging utilizing EEG & EOG data

Description: This project aims to develop an automated system for sleep disorder diagnosis and sleep stage classification using Electroencephalography (EEG) and Electrooculography (EOG) signals. By analyzing physiological data collected during sleep, the system seeks to accurately identify different sleep stages (Wake, N1, N2, N3, and REM) and detect potential sleep disorders such as insomnia, sleep apnea, and narcolepsy. The project leverages signal processing and machine learning/deep learning techniques to improve diagnostic accuracy, reduce manual scoring efforts, and support healthcare professionals in sleep assessment and treatment planning.

Prognosis of Parkinson's disease using smart wearables

Description: This research investigates the use of multimodal wearable sensing technologies and advanced artificial intelligence techniques for long-term monitoring and prognosis of Parkinson's disease. By analyzing movement, gait, tremor, and activity-related biomarkers extracted from wearable devices, the study aims to model disease progression, predict clinical outcomes, and support remote, objective assessment of patient health. The work contributes toward scalable digital health solutions that enable continuous care and precision medicine for neurodegenerative disorders.