Call for Papers

The Efficient Medical AI (EMA4MICCAI) Workshop addresses the critical need for computationally efficient and resource-conscious AI in medical applications. Despite remarkable progress in medical AI, real-world clinical adoption remains hindered by high computational costs, annotation burden, and deployment constraints. EMA brings together researchers, engineers, and clinicians to advance innovative strategies that enhance the efficiency, scalability, and accessibility of medical AI — without compromising performance or reliability.

Building on the success of EMA 2025, this second edition expands its focus to efficient training and adaptation of large-scale foundation models, data-centric efficiency strategies, and multimodal learning — directly addressing the growing tension between model capability and resource constraints in clinical environments.

Where to Submit

Topics of Interest

The EMA workshop will cover, but is not limited to, the following topics:

🔧 Model Efficiency

• Lightweight and low-latency architectures for real-time medical imaging and video analysis

• Pruning, quantization, distillation, and low-rank adaptation techniques for compressing medical AI and foundation models

• Emerging hardware, systems, and co-design solutions for efficient medical AI computing and deployment

🧠 Foundation Models & Multimodal Learning

• Efficient multimodal learning and foundation model adaptation: cross-modal training and alignment, vision-language models, multimodal retrieval and reasoning, and integration of diverse data sources including imaging, text, wearable data, etc.

• Efficient generative models and diffusion-based approaches for medical image synthesis, augmentation, and simulation

📊 Data-Centric Efficiency

• Data-centric efficiency: data selection and curriculum design, dataset pruning and condensation, active and continual learning, and synthetic data for label-efficient medical AI

• Scalable and privacy-preserving federated or distributed learning for clinical data across institutions

🏥 Clinical Deployment & Reliability

• High-performance and energy-aware AI for resource-constrained clinical environments, including edge devices and embedded systems

• Efficient uncertainty quantification and calibration methods for reliable clinical decision support

• Efficiency-aware benchmarking: evaluation frameworks that jointly assess accuracy, latency, memory, energy consumption, and clinical utility

Important Dates

Submission Instruction