Browsing Anatomical Image Databases: A Case Study of the Visible Human

Chris North, Flip Korn
Human-Computer Interaction Laboratory and
Department of Computer Science
University of Maryland, College Park, MD 20742 USA
+1-301-405-2725
{north, flip}@cs.umd.edu

To appear in Video Proc. ACM CHI '96 Conf, ACM Press, 1996.
(Video also available in HCIL Video Reports 1995.)

Copyright on this material is held by the author(s).

ABSTRACT

This video demonstrates two user interface prototypes for browsing the National Library of Medicine Visible Human dataset on the internet. The first uses a graphical approach and demonstrates a general interface for exploring volumetric data. The second uses a textual approach for exploring hierarchical information containing inter-relationships.

Keywords

User interface, information exploration, digital library, medical imaging, volume visualization, hierarchical information, network access

INTRODUCTION

The U.S. National Library of Medicine (NLM), for its Visible Human Project [1], is in the process of creating a large digital library of anatomical images of both a male and female subject. It contains MRI and CT scans as well as cryosection images (digital color photos of axial cross- sections). It is available via the internet to a large community of users with varying backgrounds and expertise. However, since the dataset is extremely large (15 Gigabytes for the male) and could take weeks to download, it is important for users to be able to retrieve only their desired subset of images. We propose two user interface prototypes to assist users in browsing the dataset and downloading desired images.

A GRAPHICAL APPROACH

The first prototype [3] utilizes users' knowledge of the location and visual appearance of anatomical structures within the body. It allows the user to browse a miniature version of the Visible Human digital body in order to select and retrieve images. The direct manipulation interface (Figure 1) presents the user with a coordinated pair of orthogonal 2D cross-section views of the body. The left view-window displays a coronal section (a front-view longitudinal cut of the body), which gives an overview of the dataset. The right view-window displays an axial section (a cross-section perpendicular to the longitudinal axis of the body), which acts as a preview of the higher resolution cryosection images in the dataset.

A horizontal indicator on each view indicates the position, on the body, of the cut shown in the other view. The user can vertically drag each indicator to sweep the cut, shown in the other view, through the body. The views provide smooth, rapid feedback (approximately 20 fps on a SUN SparcStation 1+) reflecting the cross-section at the sliding cut plane, resulting in a dynamic animated effect of motion through the body [4]. This allows the user to easily explore the contents of the entire dataset. When the user locates an axial section for which high resolution data is desired, the system downloads the corresponding full-size cryosection image from the NLM archive over the internet and displays it in a detail view.

Figure 1: The Graphical Prototype

A TEXTUAL APPROACH

The second prototype utilizes users' knowledge of medical terminology of human anatomy. Users can interact with portions of the Medical Subject Headings (MeSH) Tree [2] in order to select and retrieve Visible Human images. MeSH, published by NLM, is a hierarchy of medical terms pertaining to anatomy, histology, pathology, etc. Dynamic interaction with the MeSH reveals inter-relationships between terms, such as spatial proximity of anatomical structures within the body.

The browser (Figure 2) displays the MeSH tree on the left side of the screen as a table of contents that can be dynamically expanded and contracted, level by level, via a slider widget. The children of any term can be viewed separately in subtree windows which appear in the middle of the screen. A multi-selection widget filters these terms by body region. Inter-relationships can be visualized by clicking on a term, causing related terms in other windows to be highlighted. Double-clicking on any term shows a range of thumbnail images described by that term. Clicking on a thumbnail will retrieve the corresponding full-size cryosection image from the Visible Human dataset.

Figure 2: The Textual Prototype

For future work, we believe that combining these graphical and textual approaches would provide a rich comprehensive browsing environment.

ACKNOWLEDGMENTS

This research was supported by the National Library of Medicine Fellowship Program. Support was administered by the Oak Ridge Institute for Science and Education. We greatly appreciate the guidance and encouragement received from Catherine Plaisant and Ben Shneiderman.

The prototype interface software for the graphical approach is fully functional and freely available. For information, see http://www.cs.umd.edu/projects/hcil/Research/1995/vhe.html or anonymous ftp.cs.umd.edu in /pub/hcil/Demos/VHP.

REFERENCES

National Library of Medicine Long Range Plan: Electronic Imaging. NIH Publication No. 90-2197, US Dept. of Health and Human Services (April 1990).
National Library of Medicine, Bethesda, MD. MeSH -- Tree Structures, 1995.
North, C., Shneiderman, B., Plaisant, C. User Controlled Overviews of an Image Library: A Case Study of the Visible Human. Proc. ACM Digital Libraries '96 Conf, ACM Press, 1996.
Shneiderman, B. Dynamic Queries for Visual Information Seeking. IEEE Software, Nov 1994, 70-77. (Also available as CS-TR-3022, HCIL-TR-93-01.)

Browsing Anatomical Image Databases: A Case Study of the Visible Human / north@cs.umd.edu

Web Accessibility