User interfaces

 


[ Larger ]
Medical data

Circa 2013

While people are familiar with CT and MRI images, and sometimes volumes, additional data may be derived from such imagery to build 3D surface models of internal organs. Such segmented data can be pieced together to create complex annotated models of the internal structure of a subject.

The image here are some of the windows and controls for a visualization tool showing segmented data. The tool could also show the original CT or MRI data in a point cloud to correlate the segmented data with the original data.

This project was funded by the National Science Foundation. Development was in C++, OpenGL, and Qt.


[ Larger ]
Geophysics

Circa 2008-2012

Geophysics visualization correlates data from multiple sources, including gravity maps, magnetics maps, physics models, fault line data, and historical information on earthquake epicenters. Together these give clues to the underlying structure of the earth.

The image here is from a visualization tool that plotted dozens of different types of geophysics data registered into the same coordinate space. Physics model data could be isosurfaced and sliced by cutting planes. Multiple data layers could be managed to switch between alternate models, turn on and off visual elements, or add satellite images and other maps to help correlate model features with real terrain.

This project was funded by the National Science Foundation. Development was in Java/Swing, Java OpenGL, and NASA WorldWind. Windows and Mac OS X were supported.


[ Larger ]
Airline routes and spread of infection

Circa 2011

Infections spread as people move from place to place, such as by airline. Visualizing airline routes and correlating them with the spread of infection in simulation data can help researchers understand how to control infection spread.

The image here is from a visualization tool that plotted airline routes involved in a simulated infection spread. The route endpoints were geolocated and connected by arcs color and size-coded with selectable infection data.

This project was funded by the National Institutes of Health. Development was in Java/Swing, Java OpenGL, and NASA WorldWind. Windows and Mac OS X were supported.


[ Larger ]
Semi-geolocated spread of infection

Circa 2010

The spread of infections is sensitive data, and particularly if it points to a specific source, such as a company, school, or hospital. To allow researchers to study the general characteristics of the infection spread, without revealing sensitive data, the data must be obfuscated. Simply randomizing the data destroys the characteristics worthy of study, so obfuscation must be just enough to obscure exact geolocations without affecting larger scale trends.

The image here is a selection of windows and controls for a visualization tool that obfuscated the infection spread graph. Nodes in the graph were geolocated into gross regions, not specific places. Regional connectivity patterns remained without sensitive details.

This project was funded by the National Institutes of Health. Development was in Java/Swing, Java OpenGL, and NASA WorldWind. Windows and Mac OS X were supported.


[ Larger ]
Geolocated spread of infection

Circa 2009

The spread of infections is simulated using models that include virtual populations and their daily movement to and from work, school, and stores. The output of these simulations are huge graphs annotated with the geolocation and infection status for each population member.

The image here is a selection of windows and controls for a visualization tool designed to explore the population graph and display movement of individuals and the spread of infection. Terrain models, satellite imagery, and street maps all could be laid under the data to help correlate infection hotspots with workplaces, schools, highway travel routes, etc.

This project was funded by the National Institutes of Health. Development was in Java/Swing, Java OpenGL, and NASA WorldWind. Windows and Mac OS X were supported.


[ Larger ]
Graph visualization

Circa 2008-2009

Complex graph structures are common, such as social graphs of who knows whom, idea graphs of connected notions, or graphs showing the spread of infection. When graphs grow larger than a few hundred nodes, conventional 2D visualization no longer works well. 3D graph models are needed to fit more data onto the screen without making a mess. 3D models can also reveal structures and patterns in the data.

The image here is a from a graph visualization test bed that explored many different 3D graph models and their performance with very large graphs.

This project was funded by the National Institutes of Health. Development was in Java/Swing and Java OpenGL. Windows and Mac OS X were supported.


[ Larger ]
Cell signaling pathways

Circa 2004-2006

For many science applications, the user must be able to burrow deep into the data. High-level windows may provide diagrams, plots, and lists of major components. Double-click an item in a list to get the next level deeper, and double-click on it's attributes to go even further. Traditional GUI rules say nothing should ever be more than three clicks down from the top level. However, in science applications the data hierarchy is often deep and well-understood by the user. The GUI must reflect the user's expectations, so it must have the same hierarchy of levels as does the data.

The image here is a selection of the many windows in a cell signaling pathway graph visualization application. There are windows for chemical compounds, interactions, and pathways, along with listing windows and database search dialogs for all of these.

This project was funded by the National Science Foundation. Development was in Java/Swing and Java OpenGL. Windows and Mac OS X were supported.


[ Larger ]
Volume editor

Circa 1998-1999

Volume rendering focuses on making imagery from 3D grids (volumes) of data. This volume editor generated the grid through a hierarchical scene graph of filtering functions. Each function manipulated input data to generate output data, that could be filtered by another function and so on to create a volumetric scene. Volume functions could apply filtering, thresholding, color assignment, opacity control, or even add turbulence or other shading functions.

The image here is a selection of windows for a volume editor developed for the American Museum of Natural History in NYC. The GUI uses X/Motif and SGI's Open Inventor, both of which are rarely used any more. By today's standards, the gray beveled controls of Motif look pretty clunky.

This project was funded by the American Museum of Natural History. Development was in C/C++, X/Motif, and Open Inventor.

Nadeau software consulting
Nadeau software consulting