The Digital Anatomist Information System

Structural Informatics Group
Department of Biological Structure
University of Washington
Seattle, Washington, USA

The Digital Anatomist Project is motivated by the belief that anatomy is the basis of all the biomedical sciences (including clinical medicine). Manifestations of health and disease can be regarded as attributes of anatomical structures ranging in size from molecules to body parts. Therefore our goal is to represent anatomy in a comprehensive and consistent way, which should meet the needs of all biomedical applications that require anatomical knowledge. Moreover, a logical and consistent representation of anatomy should facilitate the modeling of all biomedical information, including the patient record. We have pursued two parallel tracks for representing anatomical information: 1. The generation of graphical models derived from cadaver and clinical imaging data; and 2. Symbolic modeling of the structures and relationships that constitute the human body. Our initial work with graphical representations of anatomy provided the impetus and motivation for the National Library of Medicine to establish the Visible Human Project, and our symbolic modeling has enhanced NLM's Unified Medical Language System in order to represent deep anatomical knowledge. In collaboration with the knowledge systems group at Stanford, we have now created a very large knowledge base which provides the foundation for the machine-based intelligence needed to remotely interact with biomedical image data. Our group is unique in that we are pursuing graphical and symbolic representations of anatomy in parallel and then integrate graphical and symbolic models. We believe that such an 'intelligent' anatomical model is a requirement for next-generation applications in biomedical research and clinical medicine. Our results to date give credence to an ambitious and long term vision which we are committed to realizing.

Last updated 1/26/99 by jfb