For example, in different parts of the United States the word khakis (pronounced kah-kees) can be used alternatively for slacks. While in another geographic area of the country, accents make the kah-kees sound resemble “car keys.”
Developers saw this as a fundamental problem; speech recognition software typically resembles barcode readers looking for precise pattern recognition.
Kopin developed a zeno-linguistic mutating algorithm, says Jacobsen, which provides approximately a 10,000-word vocabulary for each language Golden-i recognizes. Its speech recognition typically searches for 25 to 75 specific voice commands for a given computer screen. Golden-i listens to the user’s pronunciation of vowels, consonants and syllables comparing spoken commands with specific available screen commands. Since no two commands are identical, Golden-i selects the appropriate response requested and the mutating algorithm adjusts vocabulary with the users pronunciation.
Golden-i can presently speak and recognize English, French, German, Spanish, Japanese, Mandarin Chinese and Korean, and can switch by command. Other languages such as Hebrew, Arabic, Portuguese and Russian will be added soon.
What can be done
Golden-i has almost limitless potential in what it can do — truncated only to the innovation of software and technology.
By design Golden-i is a full PC with multiple wireless radios, six-axis head tracking, GPS and a digital compass. Miniaturized on-board accessories can be added, like a HD camera, laser rangefinders and more. It achieves its size and weight with high speed processing by leveraging the high performance systems, computers and wireless networks.
“All we sought to do was to augment and enhance people’s natural capabilities that makes them more efficient, more productive, improves their safety and leverages existing legacy systems and software efficiently” says Jacobsen.