"Doubted about the originality of an image, get it checked with Fake Finder Algorithm."
Fake Finder is an algorithm based project that has been developed at the Dartmouth college, with the developing team including James O’Brien, Hany Farid and Eric Kee. This algorithm uses the shadow data from any image or photo to analyze whether the shadows are consisted with a single source of light or multiple directional, and so concluding whether the picture or image is an original or fake(edited).

The algorithm makes use of the SIGGRAPH 2013 method, that is used to identify and calculate the following features of any shadow –

1) Length – The length of the shadow casted could help in identifying the angle of the light source.
2) Width – The width of the shadow could help in identifying the distance of the light source from the object.
3) Object Area Covered – The amount of object’s self-area that has been covered by darkness(shadow) from the angle of the light source.
4) Intensity – The amount of intensity of the shadow that has been created due to the presence of the light sources.

The steps that are included while analyzing a picture with this Fake Finder algorithm, are as follows –

1) The picture or the image is scanned in the highest possible resolution.
2) Then the scanned image is passed to the algorithm to be analyzed.
3) The algorithm then starts analyzing the shadow parts and the sources of light.
4) After analyzing, then the algorithm starts picking each set of data (light source and shadow) from every angle, and starts checking recursively whether the required calculations are matched or not.
5) If all the calculations are correct, the algorithm declares the image as “Not Fake” or else “Fake” in case of any mismatch in the calculation results.

This is one of the simplest algorithms that is used by several online websites to filter the images that were being uploaded from any other source, for validation.

Hence, the capability of the Fake Finder algorithm is still limited as there are various types of 3-Dimensional and visual editing tools available that can be used to add light source to any picture or image, and make them look like real. This leads to the negative ability of this algorithm even to declare such images as Not Fake.

• Sumant Bag
• 18 Tags
•