See photos of camera origin

A new technique will help find a digital image that is compatible with the camera used to capture it, just like finding a bullet that matches a gun. This technique will help investigators in need of evidence to convict illegal digital photos (like child pornography) c

Picture 1 of See photos of camera origin
A new technique will help find a digital image that is compatible with the camera used to capture it, just like finding a bullet that matches a gun.

This technique will help investigators in need of evidence to convict illegal suspects (such as child pornography) from a person's camera.

" When a suspect is caught with photos on a computer, and he owns a digital camera, the usual excuse is usually those photos from processing computers or downloading from the Internet ", Dr. Jessica Fridrich, assistant professor of electronics and computer mechanics at New York State University in Binghamton, said.

With the new technique, it is possible to accurately identify these photographs as the offspring of the camera that came with the suspect.

In the past, forensic scientists could find out the connection between printed photos with analog technology cameras. For example, they can compare the typical scratches on negatives with the mechanical parts of that film camera.

But so far, they have not found any advanced, cheap and reliable method to identify sophisticated scratches on digital photos and videos randomly generated by a specific camera.

It turns out that these scratches - created by the dust on the lens, the interference of optical elements and natural vibrations between pixels - are recorded on the camera's image sensor as a unique pattern of noise.

Collect noise

To determine the "noise pattern" created by a specific digital camera, Fridrich and his colleagues chose 320 photos created from 9 cameras. They upload the photos to the computer and use special software to analyze the photos at each pixel.

By assigning values ​​to vibrations between pixels, noise patterns, Fridrich created a unique "digital fingerprint" for each camera.

Later, the team used mathematical correlations to compare digital fingerprints with noise patterns from thousands of other images also taken by those nine cameras. The larger the correlation, the more likely it is that the image is created by a specific camera.

In forensic laboratory

To use this technique in real situations, experts will take a group of digital photos of the questioning camera to create its "digital fingerprint".

Next, they isolated digital fingerprints of illegal photos and compared them with the camera's digital fingerprint. The bigger the coincidence, the more likely the evidence of conviction.

In Fridrich's experiments, several thousand photos were correctly identified with the camera without any confusion, even if the image was compressed or resized.

T. An

Update 11 December 2018
« PREV
NEXT »
Category

Technology

Life

Discover science

Medicine - Health

Event

Entertainment