The Data Science Africa Visiting Fellowship presented exciting opportunities for members of the DSA community to explore and deepen their academic and professional interests through research visits to DSA partner academic institutions. This fellowship provided individuals with the opportunity to carry out independent research, build professional links, and develop their interests in data science and artificial intelligence research and policy.
Most importantly, participants in the fellowship had a valuable opportunity to establish connections with academic and industry experts, furthering their professional and personal development.
Details about how to participate in the fellowship were published later.
Computer Vision-Based Approaches for Orthopedic Patients Monitoring
This project involved researching and developing methods to improve orthopedic patient recovery monitoring using computer vision-based software tools. The need for orthopedic services was much greater than could be satisfied by existing personnel in Kenya. Caregivers were overwhelmed by the influx of patients seeking orthopedic services, such as knee replacements, especially as lifestyle diseases exacerbated the situation. Shockingly, the existing tools, such as goniometers used by orthopedic doctors and physiotherapists to monitor the recovery of patients who underwent such procedures, were prone to human error due to manual data entry, involved uncomfortable touch, and required expensive commutes to clinics. Due to their traditional data collection and storage approaches, it was difficult to draw any intelligent analyses on the trends of patients' recovery.
This was where computer vision approaches became useful. Through a web application running a pose estimation model in the backend, the project developed a software system that could stream patient joint movements, extract key metrics for tracking recovery—such as knee flexion angles—and store that data in a database with a single click. Since data was stored in a database, it could be accessed in the future and used to provide insights into patient recovery patterns, thereby improving the quality of care for patients and equipping doctors to offer quality services efficiently.
Hosting institution: DeKUT
