Error level analysis (ELA)

• “Error level analysis (ELA) works by intentionally resaving the image at a known error rate, such as 95%, and then computing the difference between the images.If there is virtually no change, then the cell has reached its local minima for error at that quality level. However, if there is a large amount of change, then the pixels are not at their local minima and are effectively original.” 

This implemenation makes use of the Python Image Library, and libjpeg (v6.2.0-882.2). ELA is carried out at 95%. Resulting ELA images have had their brightness enhanced to further seperate differences out.
http://www.errorlevelanalysis.com/

• This free research service is public

http://fotoforensics.com/

• Photo Forensics: Detect Photoshop Manipulation with Error Level Analysis

Error Level Analysis is a forensic method to identify portions of an image with a different level of compression.
The technique could be used to determine if a picture has been digitally modified. To better understand the techniques, it’s necessary to deepen the JPEG compression technique.

JPEG (Joint Photographic Experts Group) is a method of lossy compression for digital images.

The Image Error Level Analyzer
The Image Error Level Analyzer in an online tool that implements an ELA algorithm. By using it, it’s possible to rapidly discover image manipulation. The web tool is based on the Python Image Library and the libjpeg library (v6.2.0-822.2). The verification process consists of successive resaves of the image at a predefined quality. The resulting picture is compared with the original one.

The rainbowing technique
Rainbowing indicates the visible separation between the luminance and chrominance channels as blue,purple and red.
Picture modification with commercial tools such as Photoshop or Gimp can introduce distinct rainbowing pattern surfaces that have near-uniform coloring
High-quality camera photos may also include a rainbowing effect along uniformly colored surfaces.
Beware that the presence of rainbowing may only mean that an Adobe product, like Photoshop or Lightroom, was used to save the image. It may not represent proof of an intentional image alteration.
https://resources.infosecinstitute.com/error-level-analysis-detect-image-manipulation/#gref

• This is the home of Pillow, the friendly PIL fork. PIL is the Python Imaging Library. If you have ever worried or wondered about the future of PIL, please stop. We're here to save the day.

https://python-pillow.org/

• To me, the "realistic" claims for these images does not match the quality. (I've been seeing altered and computer-generated images for years. Many of those other approaches are much better than this.) Instead, the impressive parts are the speed, size, and detail. Each time you reload a picture on their site, they create a brand new image from scratch. There is no 3D modeling or manual editing. With more training and some rule adjustments, I'm sure they could fix their aesthetics issues. And they can probably do this without any impact to their post-training speed.

However, I also expect to see these images being used for scams that involve images. For example, catfishing uses photos of people to lure victims. Many catfishers grab photos from Facebook or other social media sites. But there is always the risk of being discovered through a reverse image search. With these completely computer-generated people, there is no risk of finding a stolen social media photo.
Fake dating sites could also use this system. Ashley Madison had thousands of fictional accounts for female users; 99% of their profiles for women were fake and designed to attract paying male users. Most of their fake profiles used low quality or distorted images of women
I'm actually a little surprised that I haven't yet seen fake IDs with these pictures. But it's only a matter of time before they show up. Fake accounts on services like Facebook and Twitter work best when there's a profile image, and these pictures are good enough for that. LinkedIn requires "real photos", but I'm sure these fakes would be acceptable. Really, these pictures can be used almost anywhere a headshot could be used
http://www.hackerfactor.com/blog/