Government & Public Safety
An investigation now runs on material that may have been produced by software: a voice on a call, a video from a phone, an account that looks like a person with a history. FaceOff checks voice, video and online identity and returns graded findings with the reasoning attached, so an investigator can act on them now and still explain them later
Overview
Verify identity on live camera networks, search recorded footage, and check whether a video, a voice or an online account is genuine.
An investigation has always leaned on recordings. A voice on a call, footage from a camera, a photograph sent between two phones. Whatever else was disputed, the recording itself was usually the fixed point everything else got measured against.
That is no longer safe to assume. A voice can be copied from a short sample, a face can be placed onto somebody else's body, and an account with years of apparent history behind it can be built in an afternoon. The damage runs in both directions: a fabricated recording can send an enquiry in the wrong direction for months, and a genuine one can be brushed aside by a suspect simply claiming a machine made it.
The checks below examine the material an investigation actually collects and return a graded finding with the reasoning behind it, showing what was noticed and where. That is what lets a line of enquiry be justified at the time and the work be reviewed later by a supervisor, an expert instructed by the defence, or a court.
What you deploy
These are the products investigation teams deploy most often, following a case from the material that comes in, to the identities and accounts behind it, to the cameras that might place somebody somewhere.
Autonomous voice interviewer with trimodal audio, visual and linguistic forensics.
ExploreLive face recognition across your camera network, matched against your watch-lists.
ExploreKey capabilities
What this sector asks for most often. Each one runs at the moment a decision is being made, and each returns a result a reviewer can explain to somebody else.
Recorded audio is examined for speech that was cloned or generated rather than spoken by a person, and the passages that raised concern are kept so an investigator can listen to them instead of taking a summary on trust.
Video collected during an investigation is examined for face swaps and generated footage, and the finding points to the moments that need review. A long recording becomes a short list of passages worth an investigator's time rather than a day of playback.
When somebody is impersonated, whether a suspect, a witness or an officer, the material rarely stays in one place. Suspected fakes of a named person are traced across online platforms and each match is checked against the real person, with the monitoring held as a case file rather than a folder of screenshots.
Public information about the accounts and people involved is drawn into one structured view, including profiles that were built rather than grown and accounts that move together. An investigator starts from a picture of the case instead of assembling one by hand.
Identity records held across different authorised sources can be resolved and linked, with the relationships laid out for an analyst to review. An identity assembled from pieces of several real people becomes visible as a group rather than passing as a set of unrelated individuals.
Faces on live cameras are matched against watch lists your own investigators build, and every match arrives graded for an officer to accept or reject. It runs on hardware you own inside your own network, and every action taken is recorded.
Where it is used
Every entry names a product on this site and describes it the way its own page does.
recorded voice checked for cloning
video evidence checked for manipulation
fakes of a person traced online
faces matched on live camera networks
public information turned into a case view
Common questions
A graded finding with the reasoning behind it, showing what was noticed and where in the material. It is written to be reviewed and challenged rather than accepted, because a conclusion an investigator cannot explain will not survive a supervisor, let alone a court.
DeepAudioGuard checks recorded audio for speech that was generated, cloned or altered rather than spoken, and keeps the passages that raised concern so they can be listened to. The result is graded rather than a bare yes or no.
Admissibility is decided by the court, under the rules of evidence that apply where the case is heard, and no supplier can promise it in advance. What is provided is a finding with its reasoning attached and a record of what was examined, which is what a supervisor, an expert instructed by the defence or a court needs in order to review the work.
No. A camera match arrives as a graded alert an officer accepts or rejects, and every finding reaches a person with the reasoning attached. What gets recorded and acted on is the officer's decision, not the system's.
CrossCam PersonTrack runs on your own hardware and makes no call outside your network in normal running, and analysis can run inside your own environment where case material is sensitive or must stay in the country.
Yes. Actions are recorded and can be searched, which is what allows a line of enquiry to be reconstructed months later and what shows that a decision was taken by an officer rather than by a system.
DeepFakeFinder traces suspected fakes of a named person across online platforms, verifies each match against the real person and holds the monitoring as a case file. That turns a single report into a picture of how far the material has spread, and gives the team something to take to a platform.
The checks are meant for recordings as they arrive rather than clean studio material. Where audio is difficult, VoiceExtract AI separates speakers and removes background noise and interference, so a recording that was going to be set aside can become usable.
Yes. Any method of detection can be mistaken in either direction, human judgement included, and in an investigation both kinds of error land on real people. A finding is a reason to look rather than a conclusion, which is why the reasoning travels with it and why every decision stays with an officer.
Bring your own footage. We will run it through the Adaptive Cognito Engine and walk you through the findings.
