CHPC News, November 2000, Volume 12, Number 1
by Stefano Foresti, Julio Bermudez, Dwayne Westenskow and Jim Agutter, Center for the Representation of Multi-Dimensional Information (CROMDI) University Of Utah, www.infoviz.utah.edu
The research group of the Center for the Representation of Multi-Dimensional Information (CROMDI) has developed a new visualization technology (IntuInfo+) that facilitates the rapid and accurate analysis of large quantities of complex and quickly changing data.
IntuInfo+ makes use of the human intuition of 3D geometry and visual-auditory clues that everybody understands as a result of dealing with the real world. The technology maps data variables into 3D objects and their properties. As data changes in time so do the objects. The audio-visualization permits its users to intuitively detect any change in the geometry, position, size, color, sounds, etc. of the object(s) and determine its implication. This enables significantly faster, more accurate and less cognitive demanding recognition of events than is possible using existing visualization technologies.
IntuInfo+ organizes measured or computed information into relevant data sets or critical functions. These data sets are then mapped as three-dimensional objects (e.g., cubes, spheres, pyramids, prisms) that work as metaphors of the critical functions of the system. The 3D objects' location and movement in space as well as their geometric and aesthetic attributes (e.g., shape, texture, opacity, color, etc) correspond to pertinent data variables. The addition of sound provides further perceptive acuity.
The methodology and technology of IntuInfo+ are unique, involving a multi-disciplinary, concerted, iterative effort with experts in architecture, communication, computer science, mathematics, music and psychology, and the fields of application: anesthesia, finance, etc.). Intrinsically non-spatial, non-dimensional, abstract information are visualized with animated 3D objects/spaces: each monitored variable (or functional relationship) is mapped to those objects and their attributes (such as color, texture, sounds, shape). Such mapping is studied to suit the user's mental model, utilizing intuitive understanding of 3D geometry and visual and auditory clues. The visualization provides a "health" state of a process, system or organization (health being defined as normal or expected behavior): deviations from the health state can be detected, diagnosed and/or treated. It provides for both global (history and trend) and local (moment and details) representations of the data in real time, in addition to seamlessly matching discrete (e.g., heartbeat) and continuous (e.g., respiratory volume) data.
Current methods for presenting information have limitations when dealing with large amounts of information, be they waveforms, pie charts, diagrams, icons, matrices, etc. As a result, traditional tools often produce information overload and confusion instead of insight and decision-making power. Most attempts at creating new graphic representations (e.g., 3-D computer graphics and visualization) are limited to simulating phenomena of spatial objects and do not address abstract and/or non-spatial data.
Fig. I: Traditional monitoring system
Anesthesia monitoring application
This project seeks to develop new displays for visually representing physiologic variables, to enhance a clinician's ability to see and rapidly respond to critical events. A digital architectural visualization of an individual's physiologic data is created in real time. It is both a probing and representational system that brings together science and art through architectural design. Supported by a 5-year grant from the NIH, the project's goal is to create a coordinated and interactive hyper-representation that articulates physiologic data in a format that is easily and quickly understood. Raw data is obtained from existing medical equipment that measures human physiological signs using non-invasive techniques. Using this data 3-D objects are created in digital space that represent physiologic changes within the body and show functional relationships that aid in the detection, diagnosis, and treatment of critical event.
While there is a clear understanding of how to represent anatomical aspects of the human body, the representational language for body functions and physiology (processes and states) are not as well developed. During anesthesia, the anesthesiologist watches over 30 interrelated variables charted as 2D waveform data displays to determine if a patient is stable and in the desired physiologic state.
These variable are continuously monitored by the attending anesthesiologist: Pulmonary Function, updated each breath (Tidal Volume, Respiratory Rate, Nitrous Oxide, Oxygen, Carbon Dioxide, and Airway Pressure), Cardiac Output, updated each heart beat (Stroke Volume, Heart Rate, Systolic Blood Pressure, Diastolic Blood Pressure, and Arterial Oxygen Saturation), Predicted Plasma, Brain, and Muscle Concentrations, updated every 2 seconds (Fentanyl, Propofol, Isoflurane, and Vecuronium), Fluid Changes (Blood Loss and Blood Infused), Urine Output, and Body Temperature.
All of these variables are interrelated and constantly in flux. For example, when air is inhaled into the lungs, pressure is exerted on the heart due its close proximity. This pressure causes the volume of blood pumped by the heart to change with each beat. The 2D representation, which depicts each variable separately, does not show relationships between disparate data.
A 3-D model of physiologic data is displayed in four interactive windows; each one designed to show certain information in detail (Refer to Fig II below). Departure from "normal" reference grids, shapes and colors helps the clinician discover change. The display structure maps each variable to a clinician's mental model, to help diagnose problems. Functional relationships link the elements of the display to help the clinicians treat problems.
Fig. II: The prototype models eight variables in real time on a single screen (Heart Rate: HR; Diastolic, Systolic and Mean Blood Pressures: DBP-SBP-MBP; Arterial Oxygen Saturation (SaO2); Respiratory Rate (RR) and inspired and expired gas content (O2 and CO2)).
The main features of the display design are shown in Figure II. The display is composed of four separate interactive windows; each one is designed to show certain information in more detail. The perspective view, shown in the lower right hand corner, provides a comprehensive, integrated and interactive view of all physiological data. Front, side and top views show how the same data appears from different vantage points much in the same way that architectural construction drawings are done to provide a more precise and quantitative image of certain aspects of the information.
The display permits the user to set these normalized reference grids, which match a patient's normal condition. The cardiac icon (red sphere) in the center of the front view grows and shrinks with each heartbeat. Its height is proportional to the cardiac volume (ml/min) and its width is proportional to heart beat period (i.e., 1 divided by heart rate). This graphic icon offers useful similarity to a beating heart. The torus object surrounding the sphere shows the expected normal values for cardiac and heart period. The position of this object on the screen is proportional to the patient's mean blood pressure (moving up is higher, moving down is lower). This graphic icon provides a metaphorical association with a working heart that aids in the intuitive understanding of the display.
The curtain in the background shows respiratory information. The variance of gray and red colors on the background shows inspired and expired gases. The height of the "curtain" is proportional to tidal volume. The width of each fold is proportional to respiratory rate. The green and gray colors show the concentrations of respiratory gases oxygen and carbon dioxide, respectively. Time moves from right to left, with present conditions at the "front" or right edge of each view. Past states remain to permit a 'historical view' of the data.
by Jonzy, CHPC Staff
The "word" of importance here is not a particular document creation program. Instead the "word" I am in reference to is your "password". You know, that string of characters you must enter after your login name so you can get access to a computer system. Your login name is not a secret, and many times is that portion of characters prior to the '@' (at) sign in your email address. However, your password should be a secret selection of characters known only by yourself. The importance of using "complete" and "difficult" to guess passwords will not only assist in ensuring your password remains a secret, it will also help keep unauthorized users off the computer system you are using.
Using the word "complete" implies the use of all, the maximum number of allowed characters in your password. Most computer systems use an eight character password. Thus, I cannot emphasize enough, the use of all eight characters in your password. Using the word "difficult" implies your password should not be one that can easily be guessed. A bad choice of password would be your login, your login spelled backwards, your full name, or any derivation there of, a password that can be socially engineered such as your significant others name, your favorite hobby, the make of vehicle you drive, or anything someone knows or can easily discover about yourself.
Good passwords cannot be found in a dictionary and contain a mix of upper and lower case characters, as well as the use of numbers and punctuation characters. Good candidates are phrases condensed to eight characters. For example: "!4tm24Ts" is a derivation of the phrase "one for the money, two for the show."
One for the money, two for the show ! 4 t m 2 4 T s
However, this password is no longer a good password due solely to its use in this example. With a little creative thought you should be able to come up with a unique password.
Protect your password. Never give your password to anyone. This includes systems administrators, your significant other, or any other person. Do not write your password on a stickit note and paste it to your computer. You need to diligently guard your password. It is also recommended you change your password on some regular basis. If you have accounts on more than one machine, you should have different passwords for each computer. This can make your computer use difficult, attempting to remember the password for each machine, but it also makes it more difficult for those attempting to exploit your username/password pair if it is ever discovered.
- Use all the characters allowed for a password.
- Do not use easily guessed passwords.
- Mix the character case, use numbers and punctuation.
- Protect your password, give it to no one.
- Don't write your password down next to your computer.
- Change your password frequently.
- Don't use the same password on different machines.
For more information regarding computer security issues, please see: http://www.uusec.utah.edu/security/
by Julia Harrison, CHPC Staff
The University of Utah Center for High Performance Computing is hosting a conference on cluster computing in the Sciences, February 8-9, 2001 at the University of Utah.
The conference will provide attendees a snapshot of cluster computing in the Sciences, and some roadmaps for the future of cluster computing. Speakers from the oil industry, computer industry, national labs and universities will share their views and experience in cluster computing.
Last year we held this conference with a particular emphasis on Geophysics. We will again hold this conference back to back with a Geophysics conference being held in the INSCC building, but have changed the title to reflect the more general scope we intend to provide.
We are in the initial planning stages, but will provide more details as they become available on our web site: http://www.chpc.utah.edu/cc
CHPC is in the middle of a Fall Presentation Series. Please see our home page at http://www.chpc.utah.edu for exact dates, times and locations of the upcoming courses. CHPC plans to present this series every year.
|Name of Presentation||Day||Time||Location|
|Introduction to CHPC||Sept. 28||2:00pm||INSCC 105|
|Introduction to Batch||Oct. 26||2:00pm||INSCC 105|
|Intro to Parallel Computing||late Nov.||TBA||INSCC 105|
|Intro to MPI||early Dec||TBA||INSCC 105|
|Intro to OpenMP||Spring||TBA||INSCC 105|
If you have suggestions on topics to add to this series, please contact Julia Harrison (firstname.lastname@example.org, 801-581-5172). We hope these presentations will provide useful information to our user community.