Biomedical Problem Solving Environments

The centerpiece of all CIBC activity is the Biomedical Problem Solving Environment, or BioPSE. This environment contains a suite of applications and integrated problem solving systems that address the needs of the collaborators and those of related fields in biomedicine. While this software represents the state of the art in computer science and the numerical and visualization techniques from many specialties, its main goal is to be easy to use, powerful, and flexible. The CIBC software should provide an infrastructure that liberates the scientist from the mundane tasks associated with most existing software tools. By being flexible and extensible, the software environment should free biomedical experts to add and apply their expertise to the science at hand.

Driving Goals: To be useful, a software architecture for biomedical applications must achieve the following goals:

• Accessibility and Usability: must satisfy a wide range of users (from software developers to biomedical scientists) who have applications spanning a wide variety of biomedical domains.
• Integration and Extensibility: must be extensible so that users can enhance it to fit their needs; must connect to and easily interact with other software systems.
• Performance and Control: must be easy to use and robust, while also efficient and powerful.

The SCIRun Software Framework
SCIRun is a general-purpose framework that implements "data flow", a natural and intuitive workflow organization that passes data from operation to operation. Each step is a module that has its own input and output and controls to adjust the operations it carries out. BioPSE is a set of modules and algorithms based on the SCIRun framework that carry out specific tasks of interest to biomedical researchers and engineers. Programs consist of networks of modules connected by data lines that read and generate data, analyze and visualize it, then optionally save the results for later access. The CIBC develops specific modules for these tasks as well as set of support mechanisms to allow users to develop their own modules.

PowerApps are specific editions of a BioPSE networks that are customized to a particular task or set of tasks. To the user, they resemble a dedicated application program and thus remove the need to understand the complexity and intricacies of the underlying BioPSE network; the user can concentrate on the problem at hand. The PowerApp is actually a layer on top of a dataflow network that includes an application-specific user interface and control options. In this way, the developer has all the power and flexibility available but presents a simplified interface to the user.

BioImage PowerApp. (above-left) The BioImage PowerApp processes and visualizes medical image volumes, i.e., sets of sequential scans from MRI, CT, or other three-dimensional imaging modalities. Users load data files in DICOM, Analyze, VFF, NRRD, and PICT formats, and can apply processing algorithms (cropping, resampling, histogram equalization) to specify and enhance regions of interest. Users can seamlessly move between two- and three-dimensional views to precisely control how different features of their data are displayed. BioImage allows users to gain both quantitative and qualitative insights into their data. BioTensor PowerApp. (above-right) The BioTensor PowerApp processes and visualizes data from a specific type of MRI modality known as diffusion weighted imaging (DWI). DWI provides noninvasive information on tissue structure that is otherwise only available by destructive histology. Specific applications include identifying white matter structures in the brain and muscle fiber organization in the heart. BioTensor can correct for common distortion artifacts, estimate tensors from the DWI data, and interatctively visualize the resulting diffusion tensor field.

BioFEM PowerApp. (above-left) The BioFEM PowerApp computes the electric field in a volume produced by a set of dipoles. BioFEM computes a solution to the bioelectric field forward problem. BioFEM also computes voltage values at electrode positions, which can be compared with values recorded via ECG or EKG. Seg3D PowerApp. (above-right) Seg3D is an OpenGL-based viewer and editor of scientific voxel data. Users can simultaneously view multiple orthogonal, two-dimensional slices of the data and create segmentation masks (labeled images) using manual drawing tools and automated segmentation algorithms. Seg3D integrates powerful higher-dimensional image processing and segmentation algorithms from the Insight Toolkit such as anisotropic blurring, edge-detection, image statistics, flood filling, morphological operations, and level-set segmentation algorithms.