A team from the Department of Electronics, Information and Bioengineering at the Politecnico di Milano, led by Dr. Andrea Moglia, has developed the first online application designed to help healthcare professionals select the most effective artificial intelligence model for generating 3D images of individual organs. This innovation aims to improve diagnostic accuracy and treatment planning by enabling precise, personalized imaging tailored to specific anatomical structures. The tool, which emerged from a study published in Information Fusion, evaluates both generalist and specialized AI models used in medical image segmentation—the process of identifying and outlining specific structures in 2D scans such as CT or MRI images to create detailed 3D reconstructions. The app is now freely accessible and designed for use by radiologists, surgeons, medical technicians, and other healthcare professionals involved in image interpretation and treatment planning. As Dr. Moglia explained, the app streamlines a previously time-consuming and trial-and-error process. “With this tool, professionals no longer need to test multiple models manually to find the one that produces the clearest, most accurate image. It significantly improves efficiency and consistency in clinical workflows.” Users can navigate the platform by selecting either a specific organ—such as the heart, liver, or individual vertebrae—or an anatomical region like the chest, neck, or abdomen. Once selected, the app displays all available AI models tested on relevant image datasets, ranked by performance. It also allows sorting by dataset size, effectiveness, and the model’s ability to detect tumors, lesions, strokes, and ischemic areas. Notably, the app includes both generalist models—trained on vast, diverse collections of medical images across the entire human body—and specialized models designed for a single organ or condition. Dr. Moglia highlighted a key finding: “Generalist models have proven in many cases to be just as effective as, or even better than, traditional task-specific models. This marks a turning point in medical AI, offering greater flexibility and scalability without sacrificing accuracy.” The tool also supports long-term strategic planning for hospitals. By analyzing annual surgical volumes and imaging needs per organ or anatomical area, institutions can make informed decisions about which AI models to adopt and integrate into their systems. The development team includes Pietro Cerveri, Luca Mainardi, and Matteo Leccardi, all from the same department at Politecnico di Milano. Their work builds on a growing body of research showing that AI can reduce human error, minimize bias, and accelerate image processing in clinical settings. This online application represents a major step forward in personalized medicine, empowering clinicians with smarter, faster, and more reliable tools to guide diagnosis and treatment. More details can be found in the study: Andrea Moglia et al., Generalist models in medical image segmentation: A survey and performance comparison with task-specific approaches, Information Fusion (2026). DOI: 10.1016/j.inffus.2025.103709.