Our project focuses on leveraging machine learning techniques to tackle the issue of identifying hand tremors. Hand tremors can signify various underlying health conditions, and timely detection is crucial for effective treatment and management.

Traditional methods for diagnosing hand tremors often involve costly medical examinations, which can be financially burdensome for patients. This financial strain can deter individuals from seeking timely medical attention, potentially delaying diagnosis and treatment.

To address this challenge, we are developing a classification model using machine learning. Our model utilizes data from 3 MPU9250 sensors positioned on the hand—wrist, between wrist and elbow, and between elbow and shoulder. This dataset comprises readings from 3-axis accelerometers, gyroscopes, and magnetometers, alongside a result column denoting '1' for shaking hand and '0' for stable hand. Employing machine learning algorithms, our aim is to accurately classify hand movements as tremor-inducing or stable, thereby offering a cost-effective diagnostic solution for hand tremors, ultimately improving patient outcomes.

• E/19/003, Harith Abeysinghe, email
• E/19/095, Lakshitha Edirisinghe, email
• E/19/163, Eshan Jayasundara, email
• E/19/236, Lahiru Manikdiwela, email
• E/19/443, Dasun Theekshana, email

• Project Repository
• Project Page
• Department of Computer Engineering
• University of Peradeniya