Voice Biometrics Partners > Vocalect
Vocalect offers their biometric engine using hosted subdialogs,
allowing your application to hand off control of the call to the
Vocalect server for processing and verification of your caller’s
voice. Your application will record the caller’s voice, and then
pass it to the Vocalect subdialog.
<subdialog name="dialogid" src="subdialogurl"
namelist="customer userid ..." method="post"
enctype="multipart/form-data" />
Here are the steps involved in getting started authenticating
callers with Vocalect.
- Apply for a Vocalect trial account
- Have a quick technical review session with Vocalect
- Start an enrollment session
- Record the caller’s voice, submitting it to a Vocallect subdialog
- Generate an enrollment from the recorded voices
- Start a verification session
- Record the caller’s voice and submit it to a Vocalect subdialog
- Check the result code to determine if verification succeeded.
To get started, download the Vocalect trial
account form and a copy of Vocalect's
Acceptable Use Policy (AUP). Review the AUP, complete the form and
scan it. Email it to trials@vocalect.com to get your account
provisioned. Vocalect will contact you for a quick (15-20 minute)
technical review session to make sure you have everything you need to
get started. After the session, Vocalect will provide you with your
account credentials.
The sample application included here is a very simple case. It
greets the caller, starts enrollment, and then verifies them. In a
real-world application you would likely have enrollment as a separate
step from verification, checking at the beginning of each call to see
if a caller was previously enrolled.
In this simple application, we’ll greet the caller with a
<prompt> and then start enrollment, kicking things
off by setting the account credentials provided by Vocalect.
We’ll start an enrollment session by calling the Vocalect
subdialog, then record the caller’s voice. The recording is
passed to the hosted AudioCheck subdialog for processing and then a
final call is made to the EnrollUser subdialog. In your production
application, you’ll likely want to make several recordings,
passing each one to AudioCheck before calling EnrollUser. Finally,
clean up your session by calling the FinishTransaction subdialog.
Once the caller is enrolled, verification begins. Verifying a
caller is a very similar process to enrolling them. We start a
verification session with our Vocalect credentials, then record the
caller’s voice. The recording is posted to the VerifySample
subdialog and Vocalect sets a flag for success or failure. The caller
is told what happened, and then we end our session with a call to the
FinishTransaction subdialog.
You can get
the sample VoiceXML application from Voxeo’s GitHub
account.
|