Turin, Italy-based Loquendo, a provider of speech technology solutions, announced the availability of a desktop tool, LVIS-Pre Forensic, designed for voice biometric confirmation of a suspect's identity. This new tool is very useful in meeting the needs of forensic criminal investigators, say officials with the company in a press release.

 

The Loquendo (News - Alert) Government and Intelligence Solutions deliver advanced voice biometric technology that can identify an individual using the unique characteristics of their voice. The company's speaker identification technologies are available to law enforcement, intelligence and homeland security agencies and can be configured for real-time or batch-mode processing. The company's technologies are language independent and channel independent.

 

According to Loquendo, LVIS stands for Loquendo Voice Investigation System. The LVIS-Pre Forensic is a stand-alone tool, which facilitates trained personnel to - create voiceprints and perform voice identification for investigatory purposes.

 

Some of the functionalities offered by LVIS-Pre Forensic include speaker separation, language identification, gender detection and speaker identification using standard audio files, regardless of the method of collection.

 

Officially launched during the ISS Prague conference, the LVIS-Pre Forensic is primarily designed to assist forensic specialists, who are supporting law enforcement and intelligence services, in the confirmation of a suspect's identity via a set of features such as Log Likelihood Ratio Calculation, Tippett Plots, Reference Population Selection, and Flexible Channel Normalization.

 

The Log Likelihood Ratio Calculation is a forensic industry standard used for describing the likelihood of a voice matching a specific subject versus a reference population. The Tippett Plots is a visual validation of the Log Likelihood Ratio. It shows the reliability of the results.

 

The Flexible Channel Normalization is used for operators to supply recordings from non-standard collection methods to recalibrate the system. More than 65 gender, channel and language reference populations have been built in Reference Population Selection and are selectable by the operator. 

 

Future releases have been planned in collaboration with forensic scientists and will include standardized reports that can be submitted by such experts as testimony in a court of law, said officials with the company in a press release.

 

Loquendo offers a complete range of speech technologies for server, embedded and desktop solutions - in 28 languages with 70 voices. Some of the other products of the company include Loquendo TTS, Loquendo ASR, and Loquendo Speaker Identification and Verification. These solutions are integrable via the Loquendo MRCP Server and VoiceXML and CCXML platform, say officials with the company.

In the recent news, TMCnet reported that Loquendo is providing voice technology for new e-health service in Spain. In precise, the company offered voice synthesis technology, developed by Indisys and Sadiel, for Spain's Junta de Andalucía Government new e-health service.