2026 ASEE Annual Conference & Exposition

Artificial intelligence (AI) is redefining how engineers analyze and interpret medical images, offering new ways to extract meaning from complex visual data. As AI-powered imaging tools—such as convolutional neural networks (CNNs), segmentation models, and vision transformers—become increasingly accessible, they also present a powerful opportunity for integrating authentic, data-rich experiences into undergraduate education. This study introduces a project-based learning (PBL) framework that positions AI-assisted medical image processing as a catalyst for developing students’ data science, analytical reasoning, and interdisciplinary problem-solving skills. The main objective is to help students understand how AI can enhance image interpretation, automate diagnostic workflows, and bridge the gap between data analysis and clinical decision making.

Within this framework, students engage in open-ended projects using real or simulated medical imaging datasets such as ultrasound, MRI, or endoscopic scans. They apply AI methods to identify regions of interest, classify tissue patterns, and visualize abnormalities through segmentation and reconstruction tasks. Rather than using AI as a “black box,” students examine how training data, model structure, and interpretability affect performance and ethical implications in healthcare contexts. This process deepens their understanding of both machine learning principles and biomedical imaging fundamentals.

The learning progression follows three stages. In Phase 1: Image Inquiry, students define clinically relevant questions and preprocess imaging data. Phase 2: AI-Assisted Processing introduces supervised and unsupervised learning models for segmentation, enhancement, and anomaly detection, supported by visual feedback and confidence metrics. Phase 3: Interpretation and Communication focuses on presenting findings through annotated images, dashboards, and short technical reports enhanced by AI-generated visual summaries.

An initial implementation is carried out in a multidisciplinary undergraduate course with 30 students. Early pilot results suggest that AI-powered medical image processing improves engagement, conceptual understanding, and metacognitive reflection. By embedding AI-driven imaging into data science education, this framework connects computation and visualization, preparing students for the next generation of data-informed healthcare engineering.