PRNU (Photo Response Non-Uniformity)
Definition
The unique pattern of pixel-level sensitivity variations in a camera sensor, used as a device fingerprint. A genuine camera-original image carries the source camera's PRNU pattern; an AI-generated image does not. PRNU matching requires a reference flat-field image from the claimed camera.
- Fingerprint Type
- Device fingerprint. Every camera sensor has a unique PRNU pattern.
- Source
- Manufacturing imperfections in the image sensor. Pixel-to-pixel sensitivity variations.
- AI-Generated Images
- Do not carry PRNU because they are not sensor-captured.
Common questions
What does PRNU reveal about a photo's origin?+
Every camera sensor has tiny manufacturing imperfections that create a unique pixel-level sensitivity pattern called PRNU. This pattern shows up in genuine photos from that camera like an invisible fingerprint. AI-generated images don't carry this pattern because they weren't captured by a physical camera.
How do forensic experts use PRNU to detect fakes?+
To confirm a photo came from a specific camera, examiners extract the PRNU pattern from the questioned image and compare it against a reference flat-field image from that camera. If the patterns match, the image is likely camera-original. A mismatch or absence of PRNU suggests manipulation or AI generation.
Why is PRNU stable enough for matching?+
PRNU variations come from defects baked into the sensor during manufacturing. Unlike noise that changes with each photo, PRNU is multiplicative and signal-dependent, meaning it persists consistently across images from the same camera, making it reliable for device identification.
Related terms
- Adversarial Attack
- In AI detection contexts, a forgery specifically constructed to defeat a trained detector by incorporating small perturbations that cause the detector's classifier...
- Automatic Signature Verification (ASV)
- Algorithmic systems that compare a questioned signature against a reference set using extracted biometric features (stroke curvature, velocity profile, pressure distribution). Designed...
- Bitrate Profile
- The pattern of bitrate variation across time in a variable-bitrate recording. A camera's rate-control algorithm allocates more bits to complex scenes and...
- C2PA (Coalition for Content Provenance and Authenticity)
- An open technical standard that embeds cryptographically signed provenance assertions into media files at the point of capture or editing. A C2PA...
- Codec Fingerprint
- The characteristic pattern of encoding decisions made by a camera's firmware when compressing video: the specific H.264 or H.265 profile and level,...
- Colour Pipeline
- The sequence of transformations applied to raw sensor data to produce the encoded video signal, including white balance, gamma curve, colour space...
- Content-Aware Fill
- An image editing feature (Photoshop from CS5; AI-driven from version 22.0) that replaces a masked region with synthesised pixel content matching the...
- DCNU (Dark Current Non-Uniformity)
- Additive noise generated by thermal activity in pixels even without light. Unlike PRNU, DCNU varies with temperature and camera gain settings, so...
- Diffusion Model
- A generative neural network architecture (Ho et al., 2020; Stable Diffusion, Rombach et al., 2022) that learns to reverse a noise-addition process...
- ESDA (Electrostatic Detection Apparatus)
- A non-destructive technique for detecting indented impressions on paper caused by writing pressure. A physical ink signature on paper produces detectable ESDA...
- Fixed-Pattern Noise (FPN)
- Any repeatable, spatially correlated noise in an image sensor, as opposed to random shot noise. PRNU and DCNU are both forms of...
- GAN (Generative Adversarial Network)
- A neural architecture introduced by Goodfellow et al. (2014) consisting of a generator and discriminator trained in opposition. The generator learns to...
Explained in these topics
- AI-Generated Documents, Deepfake Signatures and Image ManipulationThe unique pattern of pixel-level sensitivity variations in a camera sensor, used as a device fingerprint. A genuine camera-original image carries the source c...
- PRNU: Physics and the Camera FingerprintThe pixel-to-pixel variation in sensitivity of an image sensor caused by manufacturing imperfections. It is multiplicative and signal-dependent, making it stab...
- Matching Video to a Source CameraA fixed-pattern noise component caused by pixel-to-pixel variations in the sensitivity of an image sensor. PRNU is unique to each sensor and persists across al...