CVPR 2026 Workshop on Foundation Models for Medical Vision

About

Overview

The rapid growth of foundation models in various domains has been transformative, bringing unprecedented capabilities and advances in automated understanding. Medical vision, a pivotal segment of computer vision, is poised to greatly benefit from these advancements. This workshop delves into the integration and application of foundation models specific to the realm of medical imaging. We will cover state-of-the-art techniques for diverse medical data, such as echocardiogram, fundus, pathology, and radiology, as well as the practical challenges of implementing these models in clinical settings. Through expert-led sessions, interactive discussions, and international competitions, we aim to offer attendees a comprehensive understanding of the potential impact foundation models could have on the future of medical diagnostics and patient care.

Schedule

June 3 Central Time (Room 607)

Time	Speaker
8:30 - 9:00	Welcome and Opening Remarks
9:00 - 9:30	Jakob Nikolas Kather AI applications in oncology and cancer research (Video)
9:30 - 10:00	Faisal Mahmood Multimodal and Generative AI for Pathology (Video)
10:00 - 11:00	Coffee Break
11:00 - 11:30	Hoifung Poon Toward Virtual Patient: AI for Accelerating Medical Discovery (Video)
11:30 - 12:00	Pranav Rajpurkar Routing to Autonomous AI for Medicine (Video)
12:00 - 12:30	Tim Veenboer, Corentin Dancette, Ju-Yun Cheng, Zhou Jiang, Boyan Lv, Sebastian Aas CT Foundation Model Competition: Winner highlights

Presentations

Jakob Nikolas Kather

AI applications in oncology and cancer research

Artificial intelligence (AI) is rapidly transforming cancer research and oncology. This talk will highlight two key areas where AI is making a major impact. First, it will discuss how AI can extract meaningful, quantitative insights from medical data, including, but not limited to, histopathology images. AI tools can help predict cancer outcomes, treatment responses, genetic alterations, and gene expression, using data that is already routinely collected. These approaches have the potential to reduce the workload for clinicians, improve consistency in reporting, and reveal patterns that were previously hidden. Second, it will explore the emerging role of large language models (LLMs) and AI agents. LLMs can now reason, plan, and act—allowing them to perform complex, multi-step tasks with minimal human input. When used as agents, they can interact with external software or databases, and even help design drugs or recommend treatment strategies. These capabilities go beyond traditional AI systems and are opening up new possibilities in both research and clinical care.

Hoifung Poon

Toward Virtual Patient: AI for Accelerating Medical Discovery

Today, medical discovery advances one clinical trial at a time, each taking years to execute and often costing $100 million or more. As we enter the era of precision health in which we recognize that “one size doesn't fit all” and thus try to tailor treatments for each individual, continuing on today's discovery processes is clearly not sustainable. The confluence of technological advances and social policies has led to rapid digitization of multimodal, longitudinal patient journeys, such as electronic health records (EHRs), imaging, and multiomics. Our overarching research agenda lies in advancing multimodal generative AI to learn the language of patients and create a virtual patient world model as digital twin for forecasting disease progression and treatment response. This enables us to synthesize population-scale real-world evidence from hundreds of millions of patients and accelerate medical discovery through AI-powered virtual clinical trials, in deep partnerships with real-world stakeholders such as large health systems and life sciences companies.

Pranav Rajpurkar

Routing to Autonomous AI for Medicine

Recent evidence challenges a fundamental assumption in medical AI: that combining AI with physician expertise naturally leads to better outcomes. Studies show that AI assistance often fails to improve diagnostic accuracy and can even slow down clinical workflows. This talk presents an alternative vision: instead of forcing integration, we should embrace clear role separation between AI systems and physicians. Drawing from recent large-scale studies and advances in generalist medical AI systems, I will examine promising models where AI and doctors work separately but complementarily, each leveraging their unique strengths. Through practical examples and emerging evidence, I will demonstrate how this approach could transform clinical practice while maintaining the essential role of human medical expertise.