One of the keys to success as an inventor, says Ray Kurzweil, is figuring out how to time your inventions— not only with societal trends but also with emerging technologies that can make your vision a reality. Just look at his latest creation, the Kurzweil-National Federation of the Blind portable text-to-speech reader.
Market demand was a given, as long as the reader's price and function were right. In this case, the timing included the prescience to couple commercially available PDA and digital camera technologies into a single device that can capture text images and convert them to speech, yet is still affordable for the average consumer.
Kurzweil has been working with the National Federation of the Blind (NFB) for more than three decades on text-to-speech conversion devices, starting with readers the size of washing machines. In 2002, he predicted that affordable portable reader hardware would be feasible by 2006. Knowing that software development would probably take close to four years, he and his team of NFB scientists set to work.
Kurzweil pioneered optical character recognition (OCR) technology and the first flat-bed scanners, so he knew that the challenge would be taking Omni Font OCR recognition out of the controlled home environment and into the real world.
To read restaurant menus, signs, and other printed material, his team developed intelligent image-recognition software that could handle uneven illumination, three degrees of tilt and rotation, fuzzy focuses, odd angles, and other difficulties. This OCR problem was compounded by the fact that the device's users would not be able to see the image in the camera's viewfinder.
Therefore, the device provides a field of view report via speech output. The report offers information to help the user determine the orientation of the image (portrait or landscape) and adjust the distance from the target, since best results are obtained if the "document" fills more than 50% of the image zone. While the reader will automatically rotate any image it sees, it gets the best results if it is square with the document.
The NFB was involved throughout development. The group helped define "exactly what needed to be articulated" in terms of the interface between the unit and the user. Integrating a PDA and a digital camera, says Kurzweil, meant the K-NFB reader could take advantage of the tremendous price performance of consumer electronics. It costs $3495, rather than the $20,000 he estimates a custom-designed device might require.
But this approach also meant having to get inside mass-market devices not meant for OEM rework, to overwrite control functions and create new user interfaces—a real programming challenge, says Kurzweil. The other hardware design obstacle, he says, was developing a case that would allow for charging both units as well as communication between devices.
IN THE FIELD
About 500 NFB "pioneers" tested the K-NFB reader before it hit the market earlier this month at the NFB convention. It can read anything from labels and package information to pages in a book. And, it can store thousands of pages.
Feedback from the first users has been "extremely positive," but additional user experience will be used to refine the application. The units can be updated via software on SD cards. The next goal, says Kurzweil, is to enable the device to seek out print in the real-world environment—to scan the field of view for printed information, to identify text, and to guide the user to that information.
Looking to the future, which few inventors do as well as Kurzweil, the portable device will move beyond reading text to recognizing objects. "We are seriously pursuing object recognition," says Kurzweil, enabling the machine to recognize people and objects in a room and describe them to the user.
Inventors like Kurzweil must derive lots of satisfaction from their work, particularly when their creations help so many people. As Kurzweil's metamorphosis of mainstream consumer products clearly demonstrates, the world of electronic design is one big ecosphere, with the focused engineering achievements of one team picked up and carried by many other inventors.
To read my full interview with Kurzweil or to listen to a podcast of the conversation, go to www.electronicdesign.com and see Drill Deeper 12979.