We often say “no picture, no truth”, but in this era of rumors, news can be fake, and fake pictures are not difficult, forcing netizens to become Levin Hooke. However, since technology can be an accomplice to counterfeiters, it can also play a role in recognizing the authenticity of photos.
On Tuesday, Google ’s technology company Jigsaw (formerly Google Ideas) released a free tool that claims to help media professionals identify passive images, and even artificial intelligence-made images can’t escape the law.
▲ Picture from: Jigsaw
This photo counterfeit artifact is called Assembler and has 7 detectors, 5 of which were developed by university research teams in the United States and Italy. They are responsible for detecting photos processed by different types of technology, such as composite, Erase, etc., these operations may leave clues. A computer program that has studied a large number of cases can mark the changes it thinks after analyzing the image, such as abnormal color patterns, areas that have been copied and pasted multiple times, and whether the image has been used. Multiple camera models to create etc.
The other two detectors were developed by Jigsaw’s own team. One of them is designed to identify “deepfake”, which is the “AI face swap” that has been hotly debated in the past two years. The detector uses machine learning to distinguish between live images and style