Learn how TRULEO builds an Officer ID model to separate speech and how it differs from biometrics
TRULEO's body camera analytics platform includes Officer ID, a feature for separating officer speech in videos. While advanced, it differs from voice biometric systems and is not a voice biometric model.
Truleo’s Officer ID Technology
Truleo’s Officer ID enhances body camera footage analysis by isolating officer speech from other sounds and voices. Here’s how it works:
Context-Specific Application: Officer ID is designed for videos where the officer is wearing the camera, using positional and contextual data to distinguish the officer’s speech from other audio.
Separation, Not Identification: Officer ID separates the officer’s speech for accurate transcription and analysis but does not identify or authenticate individuals based on their voice.
Dependence on Camera Position: Effective only when the officer is wearing the camera. It cannot separate voices in videos where the officer is not the recording source.
How Officer ID models are generated
TRULEO Officer ID models are generated from an officer's videos from the last 30 days previous to their onboarding. TRULEO automatically identifies audio samples that:
- are between 15 and 60 seconds long
- most likely contain the officer's speech based on analysis of audio clarity and similarity of voice across their videos
- are of common officer phrasing, such as an introduction or explanation
A human being from the TRULEO success team then verifies the samples are only the officer's speech by removing any words from other speakers. The model can then be used to separate an officer's speech on future calls.
If a camera is shared across officers, an accurate Officer ID model cannot be generated as it is not a voice biometric of the individual officer.
Differences from Voice Biometric Systems
Voice biometrics, or speaker recognition, identifies or verifies individuals based on their unique vocal characteristics. Key differences include:
Purpose and Function:
Voice Biometrics: Identifies or verifies individuals based on their voiceprint, independent of context.
Officer ID: Separates speech in specific video contexts without identifying individuals.
Application Scope:
Voice Biometrics: Used in various settings, including security, and doesn’t require the person to be recorded by a specific device.
Officer ID: Specific to body camera footage and reliant on the camera being worn.
Identification Capability:
Voice Biometrics: Identifies and verifies individuals based on their voice.
Officer ID: Does not identify individuals, only separates speech within the body camera context.