We propose a scalable and data-driven approach to learn shape distributions from large databases of healthy organs. To do so, volumetric segmentation masks are embedded into common probabilistic space that is learned with variational auto-encoding network. The resulting latent representations leveraged derive zeroshot few-shot methods for abnormal detection. proposed distribution learning illus...

in order to improve the level of intelligence, availability, being on demand, convenience, information and humanization of rural public transportation systems, they are more willing to use modern information and communicative technologies. in addition to management services, intelligent transportation systems can provide passengers, drivers, travel agencies and other institutions with other ser...

Using natural language processing tools, we investigate the semantic differences in medical guidelines for three decision problems: breast cancer screening, lower back pain and hypertension management. The recommendation may cause undue variability patient treatments outcomes. Therefore, having a better understanding of their causes can contribute to discussion on possible remedies. We show tha...

The field of healthcare informatics is rapidly evolving. The new models and protocols of medical information system (MIS) are developed. Despite obvious advantages and benefits, practical application of MIS in everyday practice is slow. Cloud computing have emerged as an important new field, distinguished from conventional distributed computing by its focus on large-scale sharing, innovative ap...