Fighting crime and terrorism with voice recognition
Voice-recognition technology can play a valuable part in preventing and solving crimes, including terrorist acts. The EU-funded project SIIP is developing software to reliably identify speakers in legally intercepted phone calls and social media. The database will be accessible to enforcement authorities through Interpol.
© djvstock - fotolia.com
Police need to identify suspects quickly and accurately in their fight against crime and terrorism. Terrorists and criminals use a range of tactics to avoid recognition, and unknown speakers often take part in legally intercepted calls. Sophisticated voice-recognition technologies with a global reach are important for successfully prosecuting those involved in illegal activities.
SIIP aims to develop voice-recognition software to identify unknown speakers in communications, including people participating in phone calls and social media.
The aim is to make voice-identification technology a more effective, reliable source of intelligence and evidence in criminal investigations. Terrorist or criminal threats, including ransom, could be stopped earlier, saving both time wasted by police in chasing the wrong leads, and taxpayers’ money.
The software will fuse multiple speech analytic algorithms to analyse any speech source and channel. It will clean up voice recordings uploaded by enforcement authorities by removing unwanted background noises and will enhance the voice to produce a unique reference for a particular speaker.
SIIP will then search local and global audio databases using key identifiers such as gender, age, language and accent. It will also search social media channels to find matches with individuals not yet known to police. Privacy is respected at all levels, and multiple safeguards are included to prevent misuse of the SIIP solution.
The SIIP global database will be hosted by Interpol and will enable authorities worldwide to both search and upload voice references. Sharing information in this way will make international cooperation more effective and help to identify suspects who can then be brought to justice. The power of the tool will increase as the database grows.
The project, which ends in April 2018, carried out field tests on the software with police forces in the UK and Portugal during 2017.