Return-Path: Received: from syr.edu (syr.edu [128.230.1.49])     by mailbox.syr.edu (8.8.8/8.8.8) with ESMTP id WAA10798     for ; Sun, 26 Apr 1998 22:14:03 -0400 (EDT) Received: from wheaten.hitl.washington.edu (root@hitl-new.hitl.washington.edu [128.95.73.60] (may be forged))     by syr.edu (8.8.8/8.8.8) with ESMTP id WAA21104     for ; Sun, 26 Apr 1998 22:14:01 -0400 (EDT) Received: from 140.142.180.187 (cs110-14.u.washington.edu [140.142.180.187]) by wheaten.hitl.washington.edu (8.8.8/8.6.12) with SMTP id TAA02988 for ; Sun, 26 Apr 1998 19:13:44 -0700 (PDT) Message-ID: <3543EA7D.3820@hitl.washington.edu> Date: Sun, 26 Apr 1998 19:16:30 -0700 From: Erik Viirre X-Mailer: Mozilla 3.0 (Macintosh; I; PPC) MIME-Version: 1.0 To: Edward Lipson Subject: Re: Fwd: Important Information about KDI Grant Proposals References: <199804211139.HAA01437@mailbox.syr.edu> Content-Type: multipart/mixed; boundary="------------79D32F6F996" X-UIDL: a45d2266ee16dec5dade04344d68402b Status: RO Ed, You are too polite. I hope you can expect a higher standard than what was sent. It actually was intended for someone else. Here is the real thing (rough draft). I've attached my NSF bio as well as the word file. This should help. Please page me at 800 573-4775 to discuss. Erik Budget Needs: EV Consultant: 25% time at 100K/year     $25K Video capture cards (several)       $5000 Software           $10K Video cameras   equipment and lenses  $5000 Optics            $5000 Optical/Electronics Technician Software developer (part of Neat Tools team) Bullets on Eye tracking People with disabilities need: non invasive, functional, INEXPENSIVE devices. People with disabilities need functional assistance: adaptive displays. In an adaptive display, the system monitors performance: reading, tracking, manipulation and then adjusts parameters: font size, speed of movement to improve performance. The user monitor will also measure fatique levels, etc. to assist in performance.     Neat Tools was developed to be freely and cheaply disseminated. The Thing interface boxes have been designed to be as inexpensive as possible to allow users to have them. Corporate development of eyetracking technology has resulted in systems that function at some level but are too expensive for disabled users.     In this project, an inexpensive eyetracker, based on off the shelf components into something that can be easily acquired by disabled users will be developed. The development costs will be carried in the KDI project, thus avoiding corporate costs. The component costs should be on the order of $400 for a video capture card, $100 for a camera and  <$50 for electronic and optical components. Given I cubed’s ability to produce a A/D interface for $50 (software included) whereas corporatedly produced devices have been sold for over $10000, a price of $500-600 is perfectly reasonable.     The other difficulty beyond price with the current eye trackers is poor functionality.  There  is a difficult set up time for calilbration and subsequent drift of calibration. We will work on several approaches to reduce the need for calibration (by automated protocols, driven by a reset) and  by the use of relative measures that will provide functional controls in a mode other than gaze pointing. There will be development of specific drivers for  specific applications: word processing, arithmetic operations (spreadsheets) and Web surfing. The Neat Tools environment will allow us to integrate eye tracking with other applications, such as voice recognition, button presses etc.     Dr. Viirre has extensive experience in research on the brain’s control of eye movements, development of eye movement monitoring equipment and of software integration.  The development of eyetracking control systems must be undertaken with the understanding of what functions the brain is doing with eye movements before one can superimpose new demands. (Figure) Adaptive Interfaces for Neat Tools.     An artificial intelligence tool kit will be built for Neat Tools which will allow it to monitor the user. Physiologic parameters will be monitored, such as heart rate, breathing rate as well as specific performance measures: eye tracking parameters, button presses. These performance measures will be correlated with specific activities in any task domain to allow the system to determine performance. For example, reading rates will be measured, if they are slow or if there is a great deal of line repetition, the font size will be increased or the word spacing altered. The AI engine will be custom configurable to each users ability. If the user has low vision or unstable eye movements the control parameters will be altered for slower performance. Research will be carried out to assess more abstract conditions, such as level of frustration and this will be used again in more elaborate or abstract fashions (eg. turning off the violent video games and turn on the home work.)     Ultimately  the eyetracker and adaptive interface will be  useful in the non disabled domain as well. -- Erik Viirre M.D. Ph.D. Research Scientist Human Interface Technology Laboratory University of Washington Box 352142       "Man is still the most Extraordinary Computer of all" Seattle WA 98107-2142                     JFK (206) 616-3071 fax    543-5380 http://www.hitl.washington.edu NSF Bio sketch 4%2F98.doc KDI Eye track stuff