Implant Translates Written Words To Braille, Right On the Retina

An anonymous reader writes "For the first time, blind people could read street signs with a device that translates letters into Braille and beams the results directly onto a person's eye." According to the article, "In a trial conducted on a single patient who already used the [predecessor] device, the person was able to correctly read Braille letters up to 89 percent of the time, and most of the inaccuracy appeared when the participant misread a single letter. The user was able to read one word a second."

Re:Why not just use the letter?

[Mal-2]

Braille is a six-bit binary code. This was done largely because the previous system -- raised type being "read" by fingers -- was slow and inadequate. Whether the input comes through your fingertips or through the optic nerve matters little. If the bandwidth is low, it helps a lot of the content is pre-digitized. That's what Braille does.

Re:a trial of one

[amRadioHed]

These aren't drug trials here, you don't need a large sample size to determine probable effects. The guy is blind. If he can suddenly read after using this device we can be pretty certain the device is responsible.

The blind know braille but maybe not latin letters

[kawabago]

You can't assume blind people are familiar with the alphabet as we see it. They recognize a letter as a dot pattern instead of latin letter. It means 'a' to them be they might never know what 'a' actually looks like.

Another Great Slashdot Summary

[neoshroom]

Slashdot Summary:

"An anonymous reader writes 'For the first time, blind people could read street signs with a device that translates letters into Braille and beams the results directly onto a person's eye.'"

Actual Article:

"The technology, used primarily for patients with retinal pigmentosis which causes patients to lose the use of their retina but to still have working neurons, can take up to 10 seconds to convert a single letter and minutes to read a single word, and can only be used with words that are printed in a large font and held up close to a person's face. Street signs, for example, cannot be read. "

__

89% of the time, it works every time

[renhwa]

"In a trial conducted on a single patient who already used the Argus II device, the person was able to correctly read Braille letters up to 89 percent of the time, and most of the inaccuracy appeared when the participant misread a single letter."

...so by "the person was able to correctly read Braille letters up to 89 percent of the time," they mean that at one point during the test, they showed the guy a 9 letter word, and he got 1 letter wrong. Did the patient 1/9 letters wrong overall, or was this one rare mistake? How many words did they have the patient read? The technology is amazing, don't get me wrong, but somebody needs to tell their product testing/PR crews how to convey performance in a remotely meaningful way.