The possibilities of a 'portable eye'
Inventor says device for blind has much broader uses
When Peter Alan Smith pulls out his phone in a crowded Back Bay restaurant, there's no clue that his Nokia is by far the most expensive mobile phone in the entire place. He has about $2,400 in software loaded onto the $600 device.
But then it becomes apparent what's unique about Smith's phone: A flash goes off when he snaps a picture of the menu, and a few seconds later, his phone has translated the page of text into speech, and started reciting the options through his earpiece at a rapid clip.
Smith developed a degenerative eye disease when he was 18, and he is now legally blind. It has been about two decades since he could read a restaurant menu independently. He first heard about the phone on a podcast series called "Blind Cool Tech" and took out a low-interest loan to buy it.
"At work, I can take a picture of two different documents to figure out which is which," says Smith, who works for John Hancock. "At home, if I'm making chili, I can take a picture of a can to make sure it's the kidney beans before I open it."
The software that translates the text in high-resolution digital photos into speech is made by KNFB Reading Technologies Inc. in Wellesley Hills. It was developed by Ray Kurzweil, the local inventor who has been coming up with technological breakthroughs for the blind since the mid-1970s. But as with many of his innovations, Kurzweil plans for the software to be useful - perhaps incredibly useful - to sighted users in a few years from now.
Kurzweil released his first reading machine, developed in partnership with the National Federation for the Blind, in 1976; on the day it was unveiled, TV anchor Walter Cronkite used its speech synthesizer as he signed off the air. The device could scan printed pages, decipher the letters, and speak the words aloud. It was about the size of a washing machine and cost $50,000. Stevie Wonder bought the first production model, Kurzweil recalls.
In the decades that followed, much of the scanning and speech technology Kurzweil developed evolved into the scanners and scanning software now built into many printers and PCs. Burlington-based Nuance Communications Inc. sells several software products originally created by Kurzweil to convert printed documents into text.
In 2002, the president of the National Federation of the Blind asked Kurzweil about portable reading technology; though his reading machine had gotten smaller, it still resided on a desktop. "There's a lot of printed material that you don't want to bring back to your desk - or you can't - like a sign on a wall or a bank ATM display," Kurzweil says.
Kurzweil predicted a pocket-size reading machine was only about six years away. Kurzweil and the federation began collaborating to develop the necessary software. It had to be smart enough to interpret a photo taken at any angle, in any sort of lighting, with random images sometimes in the background.
An interim device, released in 2006, married a Canon digital camera with a personal digital assistant; it sold for $3,500. The cellphone version debuted earlier this year. It works only on a Nokia N82 phone, which features a built-in 5-megapixel camera, with flash. The camera offers spoken feedback to the user as to whether it has captured the entire page. After about 20 or 30 seconds of processing the image and turning it into text, it starts speaking. The standard phone with software sells for $2,145, but also includes a talking GPS system, and the ability to read any Web page to its user, among other features. (It's also good at identifying the denominations on printed money.)
James Gashel, vice president of business development for KNFB Reading Technologies, is also a user of the device. "I don't use it to read books," he says, "but I use it for the daily mail, business cards, and brochures. I was at a Catholic men's retreat over the weekend, and I used it to read the schedule."
Gashel says there are about 1.3 million blind people in the United States; since February, the company has sold "thousands" of the readers, he says. But the population of dyslexics is about three times larger. A new version of the mobile phone reader will soon be available that is targeted to them. "These are people who can see print, but have difficulty tracking from word to word," Gashel says. "So this new version of the software helps people whose problem is that they get lost in a series of words on a page."
Still more intriguing is how the phone might assist other users. A prototype in Kurzweil's lab is able to photograph a document in any of seven different languages and translate it into English, Kurzweil says he has demonstrated it in public appearances, taking a photo of text in French and having the phone read it in English. It sounds a bit like Douglas Adams's fictional Babel fish - a universal translator.
"We call it 'snap and translate,' " he says. How soon will it be available? "Two or three years," says Kurzweil, adding that he is talking to cellphone manufacturers. "We also have a prototype of speech to speech translation, where you can speak in one language and have it come out in another," he says. "Right now, that requires a bit more computation than a cellphone can support," though he notes that phones are getting more powerful each year.
As it turned out, Smith didn't really need to have his phone read the menu to him last month at the Parish Cafe. He had been to the restaurant several years ago and remembered eating a tasty steak sandwich. So that's what he ordered, and we spent the meal talking about his experiences running the Boston Marathon and his tandem cycling hobby. He says he uses the reader several times a day, about equally at home and at work. He told me he's amazed by what the KNBF software can do - "it's a portable eye, essentially" - but that he's hoping the cost will come down, so more blind and visually impaired people can afford it. "The cost is prohibitive," he says.
Gashel says, "I think the cost of the phone will come down as the product expands in terms of who it can reach. The bigger the customer base, the more we can bring the price down." Still, he says, $2,100 isn't that expensive when it comes to technology for the blind. He says he recently purchased a personal digital assistant that can render phone numbers and appointments in Braille, for $4,500, and also paid $5,500 for a flat-screen TV. "And I can't even see it," he quips. "But the people who come to my house seem to like it."
Scott Kirsner can be reached at email@example.com.