- Researchers recently conducted a study that allowed them to identify which smartphone took a picture
- It was able to identify the correct smartphone 99.5% of the time and only needed one picture to do so
- The method relies around on identifying the defects in a phone’s camera sensor
The cameras in smartphones today are the best they’ve ever been. That’s not to say you’ll be tossing your DSLR out of the window anytime soon, but the improvements over time have certainly been noticeable.
But, what you may not realize is that there are still flaws in your pictures. These defects are called photo-response non-uniformity (PRNU). The flaws show up due to imperfections in your camera that happen during the manufacturing process. These are the tiniest of variations that are imperceptible to the human eye and are the focus of a new study.
Researchers at the University of Buffalo conducted the study. The researchers studied over 10,000 pictures from 30 Apple iPhone 6s and ten Samsung Galaxy Note 5 smartphones. It usually takes about 50 photos identify which standalone camera took a photo. The researchers were able to use just one picture and get a 99.5 percent accurate result when testing smartphones.
The study’s lead author, Kui Ren, shared this in this paper:
Like snowflakes, no two smartphones are the same. Each device, regardless of the manufacturer or make, can be identified through a pattern of microscopic imaging flaws that are present in every picture they take. It’s kind of like matching bullets to a gun, only we’re matching photos to a smartphone camera.
The study identified the variations in each camera’s sensor to correctly identify them. Factors that researchers looked at included the smartphone cameras projecting colors brighter or darker than average. This is called pattern noise. It’s too subtle for the human eye to pick up, but special filters can identify it and each camera has its own pattern.
Researchers say that the process of identifying this pattern could lead to a new form of authentication. They envision a system where a customer supplies a bank or a retailer with a reference photograph of a QR code. The picture is analyzed to find the specific imperfections in their camera. Then, when the customer wants to withdraw money or make a purchase, they just take another picture of the QR code. If the flaws in the picture match, the transaction continues.
There is always the threat of a hack, but the researchers say that this method may prevent them. Hackers can remove the PRNU from a picture, but doing so would damage the embedded probe signal in the QR code.
This seems promising, but we see some issues that could arise from it. Phones are stolen every day. Relying on just this form of authentication alone could lead to some pretty nasty issues. Combining it with fingerprints, iris scans, or pin codes could be the way to go. It also seems like you’d need to have your phone with you at all times and set this security up every time you got a new one.