Image tampering has been around since the advent of the photographic process. In the 150 years commencing with the invention of conventional photography by Henry Fox Talbot and Louis Jacques Mande Daguerre, and before digital photography was introduced, photographers were staging images (Loch Ness monster, Cottington fairies) and/or creating seemingly real but actually false photoART (Stalin’s ‘erasures’, Abraham Lincoln / John C. Calhoun ‘mashup’ ) crafted from disparate negatives in photographic darkrooms.Digital photography and the long arm of the Internet increased the problem of photo manipulation. Image manipulation software has become inculcated into photographer’s postprocessing of photographs, offering easy access to an extensive palette of image tampering tools. Such manipulated images are now common; although usually manipulated for fun or art some photos are manipulated for political or commercial ends.In their paper “Digital image forensics: a booklet for beginners” Judith A. Redi et al describe a recent example of this:
“In July 2010 Malaysian politician Jeffrey Wong Su En claimed to have been knighted by the Queen Elizabeth II, as recognition for his contribution to the international aid organization Médecins Sans Frontières. A picture of him being awarded by the Queen of England accompanied his statement, diffused in local media (Fig. 1.a). When questioned about the award though, the British High Commission in Kuala Lumpur made clear that the name of Mr. Wong was not included in the official knighthood recipients lists, and that the picture was inconsistent with the usual protocol adopted for knighthood ceremonies. The image was finally shown to be a splicing between an original ceremony photo (Fig. 1.b) and Mr. Wong’s face, built to increase his popularity.
Despite attempts by some camera manufacturers, no authentication process for digital photographs has yet been successfully implemented. However, a range of digital image forgery detection techniques have been developed in recent years. Collectively these techniques are known as digital image forensics, a field that analyses images to determine image veracity through identifying image manipulation artifacts. Interest in digital image forensics techniques is increasing due to the potential of image manipulation to impact on medicine, justice, news reporting and the legal and accounting professions.
There are three main forms of image tampering (copy/move, splicing, and retouching), each with their own suite of forensics detection techniques.
Copy/move forgery is one of the most popular forms of tampering, in which a target region is copied from a particular location in an image and thereafter pasted at one or more locations within the same image or a different image of preferably the same scene.
These types of forgeries are detected using approximate block matching strategy.
This is a method that splits the image into many overlapping blocks and compares them to each other in respect of specific features to determine similarity.
Image splicing techniques are used to compose one image from multiple images. The ‘doctored’ images above of Jeffrey Wong Su En with Queen Elizabeth II and the Abraham Lincoln head swap are examples of this type of manipulation.The way it works is that (usually) with one image being the main image, elements of other photographs are outlined and copied, then pasted into the original image. In the example below, one photograph at a Floriade flower wall has been augmented with another family member from a second photograph.
Detecting splicing is difficult. The current focus in these cases is investigating where the changes show up best: the places where the outermost pixels of the spliced region come into contact with the original pixels in the image.
The presence of sharp edges (or changes) in this area can help reveal the fact that these image fragments came from two different photos.
Image retouching is about making slight changes in a photo for various aesthetic and commercial purposes. The retouching is mostly used to enhance or reduce the image features. The following pair of images are of a piece of art with the photograph taken with a flash, and the image resulting from eliminating the flash and slightly adjusting the hue.
Forgery detection, in case of image retouching, involves finding the enhancements, blurring, illumination and colour changing.
Enhancements may be local (usually copy/move modifications) or global (contrast enhancements affecting the entire image) and forensic investigation requires the application of an extensive range of techniques. Forgery detection may be an easy task, if the original version is available. Otherwise, with blind detection, the task may be very challenging.
All of this new power to change our images, while fun for us, has the potential to be confusing for all those who view our images. Presently, photos are often illusive electronic constructs, globally distributed at the speed of light with little context or explanation. We do have some forensic tools that allow us to investigate some aspects of digital photo manipulation, but no way of authenticating them out of the camera. Hopefully this will change.