A photo of a half-naked man and woman in a room has been trending on social media. The man in the viral photo is claimed to be President Nana Addo Dankwa Akufo-Addo. The pages and handles that shared the photos also suggested that the woman is a lady named Evelyn Serwaa Broni, who has been rumored to have had an affair with the president.
The photo has been widely shared on WhatsApp, Twitter and Facebook and has garnered a lot of comments on these platforms.
Fact-Check Ghana has verified the photo through critical analysis of the various parts of the viral image and through the assistance of photo forensics software and concludes that the photo has been manipulated in many parts.
To clone is to create a copy. In photoshop technique, cloning involves the duplication of part of an image to create uniform patterns in a photo or to remove blemishes in a photo. In many cases, especially when not done well, cloning leaves traces in a manipulated photo.
In the said photo, the background is presented as a uniform off-white wall but the right upper part of the wall (stretching from the right of the artwork hanging on the wall) shows a different colour pattern. A thin irregular ray at the top-right shows a white colour. The traces of the irregular ray of white colour exposes it from being an intentional design but an error that cannot appear in an original photo.
The varying, irregular colour pattern is further repeated on the bottom right corner of the photo, around the right arm of the lady. This time, the off-white colour is mixed with another patch of colour similar to the colour at the back of the lady.
Also, there are traces of different colours at certain portions of the wall in the room. These can be seen in the space between the man and the woman, and the extreme right part of the wall. The different colours do not appear to be natural or original but one created by cloning errors.
When a photo is modified, it can cause distortion. The distortion often takes the form of pixelation or imperfect colouring. This is a good indicator that a photo has been altered or manipulated.
It appears the editor or creator of the photo wants to present the lady in the photo as Evelyn Serwaa Broni. Photos of the lady on social media prove that she has tattoos on her back.
One of the tattoos is located at the mid-upper part of her back and looks like a cross pendant. There’s another tattoo on the right, which is indiscernible to Fact-Check Ghana. It appears the cross-like tattoo at the back of the lady in the viral photo has been fixed there making the photo pixelated.
Also, it appears the editor attempted to fix another photo on the right part of the back (closer to the arm) to represent another tattoo but there wasn’t enough space leaving the portion pixelated. As a result, the edges of the right shoulder of the lady are jagged instead of blending in with the background of the photo as seen on the left shoulder.
Akufo-Addo is noted to wear his wedding band in almost all public appearances. In the viral photo, the image that’s presented as the president appears to have the wedding band on but the part of the photo is very pixelated. The pixelation can also be seen in the area beneath the chin and around the neck. These suggest that parts of the photo have been modified.
Objects within one frame of a photo will often have the same amount of light, especially when they are very close like the man and woman in the viral photo. However, a closer look at the photo suggests that the light surrounding that man is slightly different from what is in the room.
The light around the man is slightly dimmer than what appears to be the off-white colour of the room. Again, the light surrounding the man in the photo appears rectangularly shaped. This indicates that the photo of the man may have been cut out from somewhere and fixed in the photo to create the impression that both the man and woman were in the same room together.
Forensics by Fake News Detector by InVid WeVerify
Fact-Check Ghana also ran the viral photo in the Fake News Detector software by technology company InVID-WeVerify. According to the Poynter Institute, headquarters of the International Fact-Checking Network (IFCN), the InVID-WeVerify verification plugin “is one of the most powerful tools for spotting misinformation online“. The software helps generate forensics and data on photos and videos. It analyses them to highlight distortions, forgery and manipulations.
The Fake News Detector found many such distortions in the photo that supports the aforementioned manipulations.
The team checked for compression of the photo and cloning on the software. Under compression, the software highlighted distortions in the ghost and grid (CAGI) of the viral photo. On cloning, the software identified a copy-move forgery detection (CMFD) issue.
The reports below indicate how the software explains the various items it looked out for and the corresponding images it produced.
Sofware: The JPEG Ghosts algorithm is based on recompressing the image in multiple different qualities and subtracting each of them from the original. The resulting difference images are post-processed to highlight regions that stand out and are likely to originate from a different JPEG image. Then, the images most likely to contain interesting findings are selected (i.e. those that feature localized inconsistencies). Consistent white, light-green or light-blue regions against a black background may correspond to traces of tampering, especially if they do not correspond to edges, but to entire regions.
Fact-Check Ghana: Observing from the ghost map submitted by the software, it is conclusive that the image has been tampered with.
Software: JPEG compression operates in an 8-by-8 grid, which is near-invisible but detectable. Adding or moving an object on an image may locally disrupt this grid. The GRIDS algorithm seeks such discrepancies, and highlights them locally. The algorithm produces local white/light-green or light-blue blobs where it detects grid discrepancies. Generally, the GRIDS algorithm is less distracted by textures in the image, and focuses on grid disturbances. However, it should mostly be interpreted in combination with other outputs that highlight the same regions.
Fact-Check Ghana: The map presented by the software clearly points out discrepancies in the photo, some of which have already been noted above.
Software: This is a deep learning approach to copy-move forgery detection. Dark pixels are authentic and light-coloured ones are targeting a possible copy-move inside the image.
Fact-check Ghana: The map highlights some of the areas that have already been identified as alterations in the photo.
From the evidence presented from both critical physical analysis of the viral photo and the forensic analysis, the photo has been altered in many areas. The manipulations, therefore, cast immense doubts on the veracity of the photo.
Leave a Reply