Oar

: A system that delineates 42 different head and neck OARs. It uses a stratified approach, categorizing organs as "anchor," "mid-level," or "small & hard" to improve accuracy across varying sizes.

: A specialized deep learning model for abdominal OARs , which outperformed several standard architectures like UNet and DeepLabV3+ in accurately contouring 16 different organs. : A system that delineates 42 different head and neck OARs

: A cognitive informatics model (Object-Attribute-Relation) used in AI and brain science. : A framework for identifying human activities in

: A unified model designed for thoracic (chest) OARs . It uses a shared "encoder" to understand general anatomy and multiple "decoders" to pinpoint specific organs, reducing the workload for radiation oncologists by up to 90%. categorizing organs as "anchor

: A framework for identifying human activities in real-time using sensor data.

Researchers are focused on using neural networks to "contour" or outline healthy organs near tumors to prevent radiation damage. Significant papers and systems include: