 |
|
|
|
|||||||||||||
|
||||||||||||||
 |
|
 |
Biometric Security is No Longer Science Fiction The technology behind QVoice’s "WhoIsIt!" Biometric Security and "Star Trek" Voice Print Security program is an advanced piece of software engineering, combined with an enhanced multimedia interface that uses a business theme or Star Trek theme. QVoice wanted the interface to be fun and easy to use. The basic security program consists of three main components. The inter changeable graphical interface, The Voice and Fingerprint biometric security technology that authenticates PC's, network logon, digital signatures and email encryption for outlook and outlook express. QVoice's VoiceLock! Security program was first released in March of 1994. The program was developed by QVoice scientists who have been engaged in various posts at universities and industries in Norway, UK and the US. The development team behind "WhoIsIt!", "VoiceLock!" and " Star Trek" Security has extensive knowledge and working experience in the field of Artificial Neural Networks, Dynamic Time Warping, Vector Quantization, Template Matching and Human interface engineering. An individual enrolls his or her voice in the system by providing a voice sample of 15 seconds. The system supplies a voice model for each enrolled individual. Training "WhoIsIt!" Security involves the extraction of certain features of the individual’s voice, which are necessary for his or her identification. These features include style, stress, pitch and frequency spectrum of the voice. Testing the biometric security program involves a matching process of each individual’s features (model) with these extracted from the claimed talker. The individual corresponding to the nearest model is selected. If no “accurate” matching can be obtained, the claimed talker is rejected. " WhoIsIt!" and "Star Trek" Security accepts English and foreign languages. An individual may even use different foreign languages for training and testing the security program. Once trained "WhoIsIt!" biometric Security will analyzes the characteristics of the human voice for security access to encrypted email. WhoIsIt will replace the password requirement for authenticating the use private keys to digitally sign and decrypt email and documents by utilizing existing sound board technology and attached microphone as its delivery vehicle. The voice biometric does not place the emphasis on “ what is being said,” What you say doesn't matter, the program will recognize and identify a specific user's voice in real time based on 3 seconds of sampled speech. How does voice verification work? A voice signal contains two types of information, one which represents the underlining text and the other which indicates the talker's identity. The former is essential in machine speech recognition where the object is to convert a spoken word input into a text command as used in products like Dragon Dictate and ViaVoice. The latter is used in machine speaker recognition to identify or verify the talker. Speaker authentication systems can be operated in one of two modes, namely, identification or verification. Speaker identification involves identifying whether or not the spoken word belongs to a speaker in a group of known speakers. Speaker verification, or authentication, involves verifying whether the spoken utterance belongs to the claimed speaker. "WhoIsIt!" is a state-of-the-art speaker verification systems. This breakthrough in technology extracts voice features which characterize the speaker in terms of his or her vocal tract dimensions, acoustics and excitation. These features are unique to each individual and are used as a voice imprint for verification. QVoice's voice verification technology extends beyond current commercial systems to enable multi-lingual capability. Speakers may gain access to encrypted email and digitally sign email and documents using a variety of different languages. This unique feature of this system is functional without any trade-off in system performance or efficiency. With as little as 15 to 20 seconds of training, "WhoIsIt!" can correctly verify the claimed speaker with over 99% accuracy with less than 3 seconds of spoken speech! Rejection Audio Log Voice Capture For Play Back Identification Traditional security systems use typed passwords as their access method. The Intruder can challenge this security program without the fear of compromising their identity. QVoice biometric programs require the intruder to submit a voice print sample to gain access. The Intruder must speak into the computer microphone thereby giving up their anonymity. When the Intruders voice print sample is rejected the voice sample is captured and placed in the audit log. When the authorized user returns to their computer they can access the audit log and play back the intruders rejected voice sample. The act of access is the act of betrayal. |
