AGNITIO is the pioneering Voice Biometrics Company created in Spain back in 2004 as an start up born in the Universidad Politécnica de Madrid. During the S&T demonstration the forensic tool BATVOX will be shown. The system, in use for the last 10 years by more than 100 forensic laboratories all over the world, is a complete and comprehensive interface powered by an i-vectors based engine and has the capability of producing generative LRs. Together with BATVOX some other tools will be shown and/or explained, such as SIFT, a flexible tool that allows quick voice comparisons, ASIS, the AGNITIO database solution and BS3, a speaker spotting solution designed for military and intelligence applications.
The RedDots project (https://sites.google.com/site/thereddotsproject/) is dedicated to the study of speaker recognition under conditions where test utterances are of short duration and of variable phonetic content. At the current stage, we focus on English speakers, both native and non-native, recruited worldwide. This is made possible through the use of a recording front-end consisting of an application running on mobile devices communicating with a centralized web server at the back-end. Speech recordings are collected by having speakers read text prompts displayed on the screen of the mobile devices. We aim to collect a large number of sessions from each speaker over a long time span, typically one session per week over a one year period. The corpus is expected to include rich inter-speaker and intra-speaker variations, both intrinsic and extrinsic (that is, due to recording channel and acoustic environment). The project was officially rolled out in January 29, 2015. At the time of writing, the project has recruited 89 speakers (72 male, 17 female) from 21 countries, with a total of 982 complete sessions.
Modern trends in communication between people and commercial entities are based on mutual trust. The Voice Biometry Server by Phonexia complements a comprehensive communication channel security strategy based on accurate voice analysis using Voice Biometrics. Voice Fraud Detection and/or Voice-as-a-Password use cases are supported. The Fraud Detection use case will be demonstrated in the Odyssey 2016 Show & Tell session. This is a typical use case for companies in the fields of insurance or micro-loan business. Phonexia’s fraud detection uses Voice Biometrics to recognize the same person in several calls. In this challenging text-independent scenario, recordings of spontaneous speech are used.
OLIVE (Open Language Interface for Voice Exploitation) is a live monitoring interface developed with funding from the DARPA RATS program to provide content extraction on real-world data with high accuracy. The system is field trainable and adaptable using example data from target operating conditions and is robust to noise and multi-channel conditions. The system integrates multiple different technologies like Speech Activity Detection, Language Identification, Speaker Identification, Keyword Spotting and Audio Event Detection into a single API, but is designed to be expandable to cover many classes and use cases of interest. The system can operate with file-based audio as well as with live streaming input.”
VeriSpeak is an Android application that enables the identification and enrollment of speakers for live speaker identification, using a remote server back-end. We will also demonstrate an application of this technology that allows the secure execution of spoken commands with the voice (ex: "Call my wife”) by simultaneously identifying the speaker and the command. The app thus grants access to only those users authorized to execute the given command. Both these applications perform with high accuracy using 15 seconds of enrollment data and provide easy-to-use interfaces for testing and demonstrating live speaker identification applications.
Speech Technology Center designes SIS II as a high-end application. It is intended for forensic analysis and speaker verification. The application enables audio recording and playback, noise cancellation, and a range of useful functions for easy-to-use operations and experience. The application supports plenty of known audio formats and offers a number of options for editing, visualization and analysis of speech signals. Our presentation will show how to perform audio signal analysis by using SIS II tools. SIS II includes plugin with speaker voice verification methods. In our Show&Tell demonstration we will also focus on automatic and human-assisted speaker recognition methods. We will demonstrate several examples of voice comparison methods application to forensic analysis tasks.