This algorithm detects 'fake' images, fights fraudulent scammers
A Czech startup has created an algorithm that can detect fraudulent images. Verifeyed uses machine learning to figure out whether images have been created with the help of editing software by determining how many times the image was compressed into a JPEG file. The more compression, the more likely the file was tampered with, having been decompressed, loaded into Photoshop or photo-editing software, manipulated, and recompressed.
Verifeyed can also determine whether an image was taken with a camera or phone. Since creating the product, Verifeyed has entered contracts with a bank, a media agency, and two insurance firms with the goal of stopping scam artists who use fraudulent IDs and images to commit crimes.
Cover image: SodaHead