Recently, deep learning models have achieved impressive results in medical image analysis tasks such as image reconstruction, quality enhancement, detection, segmentation, registration, and classification. Designing deep learning architectures to solve tasks of interest in healthcare has defined a new state of the art in this field. Despite this unprecedented progress, compelling open challenges remain. Morevoer, while high-performance deep learning models usually have hundreds of layers and thousands of channels, for clinical usage, one challenge is how to pursue the best performance within very limited computational budgets.

Furthermore, powerful deep learning models heavily rely on the availability of large-scale carefully-labelled (by humans with strong domain knowledge) data, which is particularly expensive to acquire in the field of biomedical imaging and health care. Real-word clinical and health care applications often require models that are able to learn with limited annotated examples, to handle noisy, errored and multiple annotations, and to continually adapt to new data without forgetting prior knowledge.

The aim of this Special Issue is to collate original research as well as review articles addressing recent advances in designing computation- and/or annotation-efficient deep learning models, and the incorporation of these technologies into medical image analysis and healthcare applications.

Potential topics include but are not limited to the following:

• Learning segmentation/classification/recognition/detection models with limited annotated examples
• Learning segmentation/classification/recognition/detection models with weak/sparse annotations
• Learning models from multiple annotation with varying quality or noise annotation
• Self-supervised, semi-supervised and unsupervised domain adaptation or transfer learning
• Fundamental innovations on how to selectively transfer (previously learned) knowledge to facilitate the learning of new tasks
• Designing light-weight models for clinical or healthcare usage
• Designing efficient models while focusing on edge platforms
• Effective and efficient interactions between computation and annotation efficiency