[visionlist] Interest level and design requirements for screen calibration device

Gislin Dagnelie gdagnelie at jhmi.edu
Wed Jul 11 15:09:58 GMT 2007


Dear colleagues:

This is a request for your interest level and comments.

Please do not respond to the list; instead respond to
gdagnelie at jhmi.edu   I will distribute a summary of reactions by the
end of July.

I am working with an electronics and software design company to create
a vision test calibration system for use in vision test laboratories,
public health screening sites, etc.

A prototype of the system was demonstrated at this year's ARVO
(abstract # 3565; see below)

Cambridge Research Systems has several options for screen calibration
equipment dedicated to vision research, including a NIST-traceable
device; these options are fairly expensive ($1895 to $11495) and
dedicated primarily to use with their particular hardware/software.

Simpler screen calibration systems are commercially available for
prices between US $80 and $250, but they are intended mostly to perform
colorimetric measures for graphic design purposes.  They do not measure
(or worry about) timing, native pixel resolution, minimum or equidistant
achromatic and chromatic contrast increments; and they do not measure
ambient (room) illumination or subject viewing distance.

Our current prototype can do most of the things; a pre-production
version could further refine such measures, depending on user interest.

Before developing a this pre-production prototype I would like to get
some feedback from potential users/customers and test developers how
they would rate different features.

For each of the features below, please give a rating from 0
(unimportant) to 4 (crucial):

1) Native screen resolution

2) Distortion (barrel / pin cushion)

3) Gamma functions (each screen primary)

4) Primary chromaticities and white point (at multiple intensities)

4a) NIST-traceable calibration of chromaticities

5) Dithering tables (low contrast increments)

6) Configurable for light projectors

7) Room illumination monitoring (brightness, chromaticity, screen
reflection)

8) Subject viewing distance monitoring

9) Subject 3-D position monitoring

10) On-line resource database (factory and user-contributed specs for
up to 500 most commonly used screens and projection systems; user
experiences; application tips; software upgrades; etc.)

FINALLY, please indicate how much you would be willing to pay for:
     a) Features 1-4:
     b) Features 1-8:
     c) All features:
     d) Annual NIST-traceable recalibration:
     e) Annual fee to maintain on-line resource database and software:

Any comments are very welcome:


Thanks in advance for your feedback

Gislin Dagnelie
Lions Vision Center
Johns Hopkins Univ

ARVO 2007, abstr # 3565: The Eye Pod: a calibration and monitoring tool
for screen-based vision research

G. Dagnelie, K.M. Kramer, G.J. Seifert, L. Yang, and G.D. Havey

Lions Vision Center, Dept of Ophthalmology, Johns Hopkins Univ,
Baltimore, MD; Advanced Medical Electronics Corporation, Maple Grove,
MN.

Purpose: Computer-driven tests on CRT and LCD screens are ubiquitous,
thanks to widely available and affordable hardware and stimulus control
software.  However, existing calibration tools for screen resolution,
pixel size, gamut, chromaticity, gamma, and frame rate tend to be
expensive, cumbersome, and/or inaccurate.   Moreover, none are designed
to monitor experimental conditions such as ambient illumination or
subject position. We developed a prototype for such a system in a phase
I STTR.
Methods: The prototype consists of a circular pod, containing three
TAOS TSL238 sensors in a 50 mm equilateral triangle to measure
horizontal and vertical pixel positions, and hence sizes, and a TAOS
TSC230 sensor array collecting red, green, blue, and clear screen
emissions through an IR-blocking filter to measure frame rate, gamma,
and gamut; and a second probe containing a MaxSonar EZ1 range finder, to
measure subject-to-screen distance, and two TAOS TSL2561 sensors with 45*
acceptance and diffusing windows, to measure directly incident and
global ambient illumination, respectively.  A Microchip 18F8722 PIC
processor, FTDI FT245R USB-to-parallel FIFO interface and a Xilinx
Spartan-II XC2S30 FPGA assure sensor control and communication with a
host PC.
Results: Our initial prototype allows pixel size, frame rate and gamma
to be measured with better than 1% accuracy. Using separate LCD and CRT
screen tables, (x,y,Y) coordinates of the screen primaries can be
determined within 2%.  The range finder measures subject-to-screen
distance within 1" (25 mm). Assistants quickly learn to operate the
system, perform calibrations and set up the monitoring unit.  We are
starting training with normally and partially sighted subjects to
perform these tasks independently.  We are designing a phase II
prototype that will improve chromaticity (1%), range-finding (* 0.25")
and ambient illumination measures, and provide Mac and Linux
compatibility.  Our current prototype will be demonstrated at the
meeting.
Conclusions: This calibration and monitoring tool should be helpful in
vision research labs in conjunction with many vision tests.  We also
expect it to enable calibrated computer-based vision testing in general
health clinics, schools, and similar settings where specialized visual
function test equipment may not be available.

Support: R41 EY017467


More information about the visionlist mailing list