|
|
|
“Anytime, Anywhere, Anybody”
AmiVoice is a next-generation voice-recognition engine built with the latest technologies. With its world-class voice-recognition technologies, it can be put to practical use in a wide range of applications for business and daily life.
From person to person our voice can change from day to day depending on the speaker’s physical and emotional state.
Reliable voice recognition requires a technology that can track these changes swiftly and precisely. With its flexibility and reliability, AmiVoice is widely recognized as an innovative interface that makes possible more down-to-earth voice-recognition systems available anytime, anywhere, and for anybody. It has been widely implemented in many systems and media.
Voice recognition is a technology that automatically recognizes
human voices using computers. In general, “voice recognition” means
comprehending the input voice and generating text strings. In voice
recognition, both linguistic and acoustic information play an
important role. There are many obstacles to accurate voice
recognition, such as tonal differences between individuals and
background noise. In other systems, it is a common practice to
register the speaker’s voice in order to insure accuracy. This
method, however, is a not convenient for users. AmiVoice needs no
prior registration (enrollment).
Now, AmiVoice can be used not only in personal dictation systems,
but also a variety of new applications, such as a voice-navigated
Internet, e-mail, or telephone applications such as telemarketing,
ticketing, order entry, and surveys.
Speech Recognition, Past and Future
AmiVoice is third-generation speaker-independent speech recognition technology that can respond flexibly to differences in pace of speech, intonations, and accents.
First generation (1960): recognition of isolated words
At this stage, only single words could be recognized. Because the recognition of isolated words is a core speech recognition technology, research into the improvement of accuracy continues to the present.
Second generation (1991): conversion of speech to text
The use of statistics to handle grammar made it possible to recognize sequences of words, which means sentences and passages. This increased the ability to convert speech to text.
Third-generation (2001): person-oriented recognition
The need for enrollment is eliminated and problems caused by differences in speaking speed and accents are solved, enabling speaker-independent voice recognition. Led to a broader voice recognition market (audience).
Fourth-generation (2008): ubiquitous
Moving towards technology that can be utilized by anyone anywhere, at any time. Led to availability of voice recognition on many different mobile terminals.
Fifth-generation (2011): ultra voice recognition
Voice recognition technology that far surpasses speech to text, providing results on a par with human capabilities. Towards the realization of a personified communication interface achieved through ontology and intelligent reasoning.
Recognizes Anyone’s Speech Instantly
Unlike conventional speech recognition systems, AmiVoice
doesn’t require any prior information about the speaker’s
particular field of knowledge or experience. It allows anyone,
irrespective of age or sex, to enjoy easy voice-recognition.
For Any Speaking Style
AmiVoice understands the differences in accents and
intonations unique to each region. With its “Continual Speech
Recognition” technique, AmiVoice recognizes people’s natural
ways of speaking.
For Any Speaking Speed
Some people speak quickly in times of emergency. Older people
tend to speak slowly. AmiVoice is flexible enough to follow
everyone’s speaking speed.
High-Quality yet High-Speed Processing
Recognition accuracy (recognition rate) is essential for a
practical voice-recognition system. When speech is clear
enough, AmiVoice recognize it in real-time. It also allows a
variety of applications.
Noise Resistant
Background noise can greatly affect voice-recognition
accuracy. Thanks to its world-class technologies, AmiVoice is
highly noise resistant, and can be used in such applications
as telephone-based voice-input as well as voice recognition
from such mobile environments as automobiles, trains, and
airplanes.
|
|
|
 |
|
|
|
