Automatic Detection of Parasites in Stool and Blood Smear Images using Trustworthy Machine Learning

We want to design trustworthy AI models for parasite diagnostics that (i) estimate and minimise prediction uncertainty by out-of-distribution generalisation, (ii) exploit the potential of self-supervised learning, (iii) embed
context to come up with general-purpose models invariant to region, patient profile, and parasite type, and (iv) only have minimal computational requirements, thus facilitating deployment on edge devices.