Felix Obete

At Uganda Christian University, I researched few-shot classification of Neglected Tropical Skin Diseases, focusing on underrepresented dark skin tones. I explored a framework combining Supervised Domain Adaptation and Supervised Contrastive Learning to address data scarcity and domain shift. My work improved baseline performance of DenseNet161 and ViT-B-16, reaching up to 73.33% accuracy in 5-way 5-shot settings. I focused on representation bias and the reliable deployment of dermatological AI in low-resource settings. At Associated Mapping Professionals, I developed and deployed the Field Star Android Mobile Mapping Application, enhancing operational efficiency for over 600 land surveyors in Uganda. I also set and led a team mapping 20,000 land lots for tenure registration for vulnerable communities. Earlier, I worked in land surveying and GNSS, including developing and maintaining spatial geodatabases and training interns.