2026 IEEE 15th Image, Video, and Multidimensional
Signal Processing Workshop
June 16-19, 2026, Shenzhen/China
Special Session #01 Artificial Intelligence for Healthcare: From Multimodal Data to Clinical Impact
Artificial Intelligence for Healthcare: From Multimodal Data to Clinical Impact
Session Organizers:
• Prof. Cheng Chen, Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong, cchen@eee.hku.hk
• Prof. Liangqiong Qu, School of Computing and Data Science, The University of Hong Kong, Hong Kong, liangqqu@hku.hk
• Prof. Lequan Yu, School of Computing and Data Science, The University of Hong Kong, Hong Kong, lqyu@hku.hk
• Prof. Shujun Wang, Department of Biomedical Engineering, The Hong Kong Polytechnic University, Hong Kong,
shu-jun.wang@polyu.edu.hk
• Prof. Lei Zhu, Robotics and Autonomous Systems (ROAS) Thrust & Data Science and Analytics (DSA) Thrust, The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China, leizhu@hkust-gz.edu.cn
• Prof. Xiaodan Zhang, College of Computer Science & Beijing Institute of Artificial Intelligence, Beijing University of Technology, Beijing, China, zhangxiaodan@bjut.edu.cn
Session Structure
The proposed special session will consist of an invited keynote talk and oral presentations.
Download: Special Session #1.pdf
The field of Artificial Intelligence (AI) for healthcare is rapidly evolving. Advances in sensing technology, data storage, and data transfer speeds, together with the widespread deployment of electronic health records, have led to an unprecedented volume of healthcare data. Medical images and videos, physiological and wearable signals, clinical text, and laboratory measurements are continuously collected in hospitals, clinics, and home-care settings. However, much of this information is still underutilized in daily practice, and clinical decision-making often relies on manual inspection and fragmented data. At the same time, healthcare systems worldwide are facing increasing pressure from ageing populations, rising costs, and workforce shortages, calling for more intelligent and automated solutions.
The combination of rich healthcare signals with recent advances in AI and machine learning is transforming how we analyze health and biomedicine data and support clinical workflows. Image, video, and multidimensional signal processing provide the foundations for building AI systems that can assist in screening, diagnosis, prognosis, treatment planning, and patient monitoring. New approaches based on deep learning, as well as large-scale foundation models and generative models, are emerging to tackle diverse tasks such as risk prediction, early warning, and personalized therapy selection. At the same time, these developments raise important challenges related to robustness, transparency, fairness, privacy, and safety that must be addressed before AI can be trusted in routine care.
This Special Session aims to host original papers and reviews on recent research advances in AI for healthcare. Welcomed submissions would fall in the following wide range of topics:
• AI-driven analysis of medical images, videos, audio, text, and multidimensional signals
• Machine learning for screening, diagnosis, prognosis, and treatment planning
• Analysis and modelling of physiological, wearable, and remote-monitoring signals
• Multimodal and foundation models combining images, signals, text, and clinical data
• Generative and reconstruction methods for healthcare data (e.g., enhancement, synthesis, de-noising)
• Intelligent agents and multi-agent systems for clinical decision support
• Robotics and embodied AI for image-guided interventions, rehabilitation, and assistive healthcare
• Uncertainty estimation, interpretability, fairness, and safety in healthcare AI
• Privacy-preserving, distributed, and federated learning for healthcare data
• Human-AI collaboration, workflow integration, and clinical deployment ofAI systems
Explainable Machine Learning: Multimodal Data, Architectures, Applications, and Beyond
Session Organizers:
• Prof. Haoran Li, School of Artificial Intelligence, Shenzhen University, China, lihr@szu.edu.cn
Session Structure
The proposed special session will consist of an invited keynote talk followed by oral and poster presentations.
Download: Special Session #2.pdf
Explainable machine learning (ML) is an AI subfield of growing importance. The deployment of machine learning agents in high-stakes scenarios, such as healthcare, industry production, finance, and social networks, calls for the transparency, interpretability, and reliability of machine learning models. Due to the black-box nature of vanilla deep neural network (DNN) models, specific algorithm designs for DNNs are required to achieve various levels of explainability required by different scenarios, ranging from user-friendly explanations to safety- related guarantees. At the same time, various types of input data modalities, including images, text, time series, and ever-evolving DNN model architectures, involve different technical challenges in designing explainable machine learning algorithms.
The evolving data modalities, model architectures, and real-world applications are raising various research questions in the explainable machine learning field. Beyond the technical challenges of explainable AI brought by single data modality and corresponding DNN architectures, the integration of explanation for multimodal data mirrors human perceptions and offers a more complete understanding of the AI system. Designation of explainability criteria, including the definition, distinction and measurement of transparency, interpretability and reliability, is required for different application scenarios. Beyond practical developments, the development of theoretical frameworks for explainable machine learning helps establish the foundation of explainability methods and under what assumptions they deliver valid, provably approximately correct answers.
This Special Session aims to host original papers and reviews on recent research advances in explainable machine learning. Welcomed submissions would fall in the following wide range of topics:
• Explainable ML methods for image, videos, audio, text, and multidimensional signals
• Interpretable ML for critical decision making
• Explainable multimodal analysis and integration of explainable AI systems on various modalities
• Explainable methods for deep learning model architectures and foundation models
• Explainable AI systems for medical information processing and clinical decision support
• Interpretable and trustworthy robotics and embodied AI
• Human-AI collaboration, workflow integration, and deployment of explainable AI systems
• Out-of-distribution detection and adversarial robustness of explainable ML system
• Privacy-preserving, distributed, and federated learning for explainable machine learning
• Criteria, measurements, and principles of explainable machine learning system design
• Theoretical framework and analysis of explainable criteria and methods