Tim,
I agree that making a new Optacon significantly better than any previous
Optacon would make the project much more doable and exciting. I believe that
your first idea, making the tactile image of all stroke widths only one pin
wide regardless of the image stroke width, is doable and my experience
suggests that this would be beneficial. There are standard image processing
algorithms (e.g., erosion, skeletonization, etc.) that can do this. In fact,
if the camera has 10 times the resolution of the R1D camera, and these
algorithms are applied to the image, it may be possible to use fewer tactile
pins/finger than the R1D, which reduces cost, since the tactile stimulators
are the most expensive part.
Speaking of cost, time has been unfavorable to Optacon technology.
Telesensory's present price of an Optacon is $4,495, which is more than a
good computer, scanner, and OCR system these days. In the early Optacon days
such an OCR system cost 10 times what an Optacon cost. While I believe that
TeleSensory has a very high margin in their price, it is still difficult to
see the price/performance value in the Optacon. If training costs are
included, the cost/benefit appears even worse.
On your second suggestion, a three-finger display, you may be right that a
simplified tactile image would increase the information throughput. Many
years ago at SRI we built a two-finger Optacon without the simplified image
you suggest. I expected reading speeds with this two finger Optacon to be
greater than with a one finger Optacon because I had gotten increases in
reading speed for every additional column in the display. I was surprised to
find no improvement with the two finger Optacon (unless both index fingers
were used).
It would be straightforward to test both of your suggestions before commiting
to a new design. Testing the "many image points-to-one pin wide lines" idea
could be done with an Optacon II, scanner, and computer. A computer program
could take high resolution images from the scanner, process them, and display
the simplified image tactually on the Optacon II using the RS232 port on the
Optacon II. Tactile arrays from three Optacons could be used to rig up an
experiment to test the three finger idea.
I have seen exceptional people under exceptional conditions read short
passages at a rate of about 100 WPM in an objective laboratory. However, I
expect that an average of normal people in normal conditions would be closer
to 30 WPM. It would be a major advance if changes could be made that would
increase this 30 WPM to 100 WPM.
But I don't believe that anything will be done unless there is organized
pressure from blind people. I believe there is a chance that this pressure
could result in funds being made available to improve the Optacon and
continue its production.
Jim Bliss
This archive was generated by hypermail 2b29 : Sun Dec 02 2012 - 01:30:03 PST