Forensic Technology

Amped Authenticate^®

Specialist software for verifying whether a photo or video is original, edited, or AI-generated — used by digital forensic experts and intelligence agencies worldwide. When a piece of media is disputed, Authenticate tells the court what it really is.

Call +61 499 475 408 Start an Enquiry

Deepfake detectionCamera identificationCourt-ready reportsAustralia-wide service

40+

Authentication tools & filters

14,000+

Camera JPEG signatures for source matching

Image + Video

JPEG, TIFF, PNG, HEIF, RAW & all video formats

Science-based

Filters built on published, peer-reviewed research

Why It Matters

"That Photo Is Fake" — Now You Can Find Out

Doctored screenshots, edited photos, and AI-generated images are increasingly raised in family law disputes, domestic violence matters, insurance claims, and workplace investigations. When one party claims an image is fabricated — or produces one that is — the question can no longer be settled by looking at it. Amped Authenticate examines the file itself: its metadata, compression history, sensor noise, and hundreds of subtle traces that editing and AI generation leave behind.

Findings are documented in a customisable forensic report with the scientific references behind each analysis — designed for legal proceedings and defensible under cross-examination, backed by our expert witness services and chain of custody controls.

Disputed evidence

A message screenshot or photo tendered in proceedings is challenged as edited or fabricated.

Deepfake concerns

Intimate images, voice or video content suspected of being AI-generated or face-swapped.

Source attribution

Establishing which camera or phone actually captured a photograph in question.

Insurance & workplace

Verifying photographic evidence in claims, disciplinary matters, and investigations.

Image Authentication

Is This Photo Original, Edited, or Generated?

Authenticate runs a structured workflow across the file, its global characteristics, and every local region of the image — building a picture of how the file was created and what has happened to it since.

File format and metadata analysis — EXIF, structure, and modification traces
JPEG quantization table matching against 14,000+ camera signatures
Multiple JPEG compression detection — evidence of re-saving after editing
Forgery localisation — mapping the exact regions of an image that were altered
Resize, rotation, and digital zoom artefact detection
Social media platform detection (Facebook, Instagram, WhatsApp and more)
Shadow and sun-position consistency checks for scene verification

Deepfake Detection

AI-Generated Face Detection with Confidence Scoring

Authenticate's Face GAN deepfake filter analyses facial imagery for the statistical fingerprints of AI generation — including output from modern diffusion tools such as Midjourney, DALL-E and Stable Diffusion. Each face examined receives a confidence score from 0 to 1 and a clear classification: GAN, Not GAN, or Uncertain.

That honest third category matters. A forensic finding that can say "uncertain" when the evidence is genuinely ambiguous is a finding a court can trust when it says "generated".

Combined with camera identification and compression analysis, this gives a multi-layered answer to the modern question: "was this image ever real?"

Under the Microscope

The Analyses Behind a Finding

No single test proves an image genuine or fake. Authenticate layers dozens of independent analyses — each grounded in published research — and a finding is only reported when the evidence converges.

Metadata & EXIF Analysis

Camera make and model, timestamps, GPS coordinates, editing software traces, and structural anomalies in how the file was written.

JPEG Quantization Matching

Every camera compresses images its own way. The file’s compression signature is checked against a database of 14,000+ camera models to test whether it matches its claimed source.

Multiple Compression Detection

An image saved, edited, and re-saved carries mathematical traces of each generation — evidence the file is not the camera original.

Social Media Detection

Identifies the processing fingerprints of Facebook, Instagram, WhatsApp, Twitter and other platforms — establishing where a file has travelled.

Clone Region Detection

Finds areas duplicated within the same image — the classic technique for removing objects or people from a scene.

Local Noise Consistency

Spliced-in content carries different sensor noise than its surroundings. Noise mapping localises the exact regions that don’t belong.

Interpolation Artefacts

Resizing, rotating, and digital zoom each leave periodic patterns in pixel data that survive attempts to cover them.

LCD Recapture Detection

Detects photographs taken of a screen — a common way of laundering an edited image back into an apparently fresh photo.

Light, Shadow & Perspective

Geometric consistency checks — including sun position verification against the claimed time and place — test whether the scene could have existed as photographed.

Get in Touch

Not Sure Where to Start?

Call for a free, no-obligation assessment of your situation. We'll tell you honestly what's possible and what it costs — before you commit to anything.

Call +61 499 475 408 Submit a Confidential Enquiry

Beyond Still Images

Video Authentication & Camera Identification

The same forensic rigour applies to moving footage and to the question of which device captured the evidence.

Video Integrity Analysis

Frame and GOP structure analysis, macroblock and motion vector examination, double-compression detection, and identification of cloned or duplicated frames — the traces left when footage is cut, re-encoded, or spliced.

Camera Identification

Every camera sensor leaves a unique noise fingerprint. Authenticate builds a Camera Reference Pattern from sample images and matches disputed photos or video against it — even when files are compressed, rotated, downscaled or cropped.

Court-Ready Reporting

Every analysis is compiled into a customisable forensic report citing the scientific literature behind each filter — built for accuracy, repeatability and reproducibility, and supported in court by our expert witness testimony.

Amped Authenticate works alongside Oxygen Forensic Detective in our lab — Oxygen extracts and preserves the evidence; Authenticate verifies what it really is. Read more about why screenshots often aren't enough in court.

The Process

How a Media Authentication Engagement Works

A structured workflow from the moment the file arrives to the moment the report is tendered.

Preserve the Original

We obtain the best available copy — ideally the original file from the device, preserved under chain of custody with hash verification before any analysis begins.

File & Global Analysis

Metadata, file structure, and compression history establish where the file came from, what software has touched it, and whether it matches its claimed source device.

Local & AI Analysis

Noise mapping, clone detection, and geometric checks localise any altered regions; facial imagery is screened for the fingerprints of AI generation.

Report & Testimony

Findings are compiled into a forensic report citing the scientific basis of every analysis — with expert witness testimony available if the matter proceeds to hearing.

Common Questions

Image & Video Authentication FAQ

Can you really tell if a photo has been edited?

Very often, yes. Editing changes more than the pixels you see — it alters compression signatures, noise patterns, and file structure in ways that are measurable even when the edit looks seamless. No reputable examiner promises a definitive answer on every file, but a layered analysis across dozens of independent tests usually produces a clear, defensible finding.

Can deepfakes and AI-generated images be detected?

Frequently, yes. AI generation leaves statistical fingerprints that specialised filters detect, including output from modern diffusion tools. Results come with a confidence score, and where the evidence is genuinely ambiguous the finding says so — which is exactly what makes a positive finding credible in court.

What do you need from me for an authentication?

The closest thing to the original file you can lawfully obtain. A photo forwarded through WhatsApp or downloaded from social media has been recompressed and stripped of metadata, which limits what analysis can show. If the original device is available, a forensic extraction preserves the file exactly as the camera wrote it — and that is the strongest starting point.

Will the findings stand up in court?

The analyses are based on published, peer-reviewed research, and every report documents the method behind each finding so it can be independently reproduced. Combined with chain of custody documentation and expert witness testimony, the work is designed from the outset to withstand cross-examination.

Someone claims a genuine photo of mine is fake. Can you help?

Yes — authentication works in both directions. Demonstrating that a file is a camera original with no traces of editing, and that it matches the device that captured it, can be just as decisive as proving a forgery.