Visible Human Project Image Processing Tools

Principal Investigator: Stephen Aylward
Funding Agency: National Library of Medicine
Agency Number: CCR-9996434

Abstract
The goal of the proposed work is to promote the use and development of segmentation and registration methods in medical imaging. We seek to establish a cross-platform, extensible, and easy-to-use software library and applications as the standard segmentation and registration toolkit (SRT) for medical image processing (IP). One of the strengths of the proposed toolkit is that it addresses the needs of novice and expert IP researchers as algorithm implementers and the needs of medical researchers as algorithm users. These often-conflicting sets of needs are independently resolved within the toolkit because it defines an application programmer's interface (API) at multiple levels, i.e., for functions, classes, and visual programming modules.

Specifically, we propose to (Aim 1) promote an SRT that embodies a multi-level API design philosophy, software library, and applications and that provides effective methods for medical image segmentation and registration, (Aim 2) ensure the broad utility and extensibility of that software by simultaneously developing and evaluating it at multiple labs and with respect to the needs of a wide range of users, and (Aim 3) promote the use of the SRT and communication among SRT users by providing a web-based system for releasing the software, collecting software extensions, and collecting and reporting results from other segmentation and registration methods applied to a standard image set.