AI-Augmented Annotation Tool for Education of Histopathology

Martin Saraka

Supervisor(s): Ing. Martin Dubovský

Slovak Technical University


Abstract: Effective teaching of complex and specialized subjects, such as histopathology, requires more than traditional methods, which often lack interactivity and personalization necessary for deep understanding. The integration of artificial intelligence (AI) into educational tools opens new possibilities for enhancing the learning process through adaptive support and interactive feedback. We present an AI-assisted annotation tool for histopathology education, designed to balance automated assistance with user control while fostering critical thinking and expert oversight. Our system incorporates manual editing, partially AI-driven suggestions, and Wizard-of-Oz simulations to explore advanced tutoring functions. Key features, including contextual textual hints, visual overlays, interactive learning cards, and curated study materials, aim to enhance diagnostic training by providing targeted guidance and reinforcing key morphological concepts. Through an iterative, user-centered design, we analyze how AI-powered supportive functions influence the accuracy and efficiency of histopathological image annotations among students and whether these features can aid laypersons in grasping fundamental histopathological principles. By evaluating the impact of these AI-driven explanations and interactive modules, we aim to identify strategies that minimize cognitive load, improve retention, and optimize the integration of AI in medical education. Our findings contribute to the development of AIenhanced educational tools that support deep learning, improve diagnostic skills, and create a scalable, adaptable framework for medical training. The system is now prepared for extensive student-based evaluation to assess its effectiveness in fostering a deeper understanding of histopathology and its broader implications for AI integration in education.
Keywords: Human-Computer Interaction
Full text:
Year: 2025