(VFM) Match the voice of a suspect to an image or perform 1:N and compare to a set of images from a database. Images with the highest probability of matching are identified. This process involves comparing the generated face signature from voice to each image in the database and determining the likelihood of a match. The images with the highest probabilities of being a match are then presented for further investigation.
Law enforcement, military, intelligence agencies and prison authorities are facing difficulties in identifying individuals behind unknown voices in large audio files.
Corsound's technology provides a solution to this problem by using existing data to determine if there is a match to the facial image of a suspect allowing organizations to quickly identify indiviuals.
Additional Unknown Data
Voice can be matched if a voiceprint exists. A face estimation is generated without reference data providing the probability of the presence of a suspect
The VFM solution identifies the images with the highest probability of being a match to the suspect, providing a shortlist of potential suspects
VFM technology allows for the comparison of a generated image to images in a database, to determine the likelihood that the images are a match based on specific features
Processing Speed of Voice Signature
Audio is processed at a speed of 1:150
Voice signatures cannot be reverse-engineered back to audio
Voice Signature Storage Options
There are two options for storing voice signatures: on-prem and in the cloud