Identifying Fake Images:
Identifying falsified images can be straightforward if you know a few tricks.
Images that we believe to be real but are in fact fake are bothersome because they unfairly manipulate our sense of truth.
If an image is deemed suspicious, then we can first look for clues by visual inspection and then proceed with scientific inspection if necessary.
A fake image created by altering the context is the hardest to positively identify as fake since the image is real and will pass scientific tests on the validity of the image itself.
Understanding the image formation properties of a camera can help to recognize fake images. The step-by-step physical process of forming an image is called the imaging chain and every image must adhere to the physics of an imaging chain, from the radiometric source to the final image product. Image chain analysis methods are used to examine images for evidence that the laws of physics have been broken. Any inconsistencies found within the image can be an indication the image has been altered.
A common inconsistency found when the image content is altered is the mismatch of radiometric or illumination conditions between the altered part and the rest of the image. The altered part of the image may have shadowing that is not consistent; indicating that is was illuminated under different conditions from rest of the image.
This is commonly seen when an object captured at one time of the day is added to an image that was captured at a different time of the day. Also, the light illuminating the altered part may not be consistent with the diffuse or specular light illuminating the rest of the scene. This effect is commonly seen when an object captured with a photographic flash is added to an image that was acquired with outdoor or studio lighting. Color, contrast, and tone will also vary for different illumination conditions, thus creating a mismatch of these characteristics between different images.
The physical traits of the image that can be assessed include the illumination conditions, edge sharpness, resolution, tone, relative scale, and noise characteristics.
Creators of fake images usually ignore the known physical properties of creating an image with a camera. The most significant camera effects are edge sharpness, influenced by the lens diffraction, focus, and motion blur; perspective geometry; and noise properties, usually from the detector and compression. Computer animated images are usually created without any camera effects since this will degrade the image quality and make the images less appealing to the audience. This, however, results in images that are physically impossible to capture with a camera in the real world.
When an object is added or deleted from an image, an edge is usually created that has a sharpness that is inconsistent with the rest of the image. Even an in-focus image will exhibit some blurring due to the diffraction of light from the camera aperture.
The behavior of the blurring in the image is well understood and can be mathematically modeled if the camera design is known. Even if the camera design is not known, measurements within the image can produce a relatively accurate mathematical model of the camera that can provide reasonable predictions. Cutting an object from one image and inserting it into another image will create a sharp edge at the boundary of the inserted object that is sharper than physically possible. This sharpness is easily seen and creates an obvious sign that the image has been altered, so smudging tools in image processing software are usually used to reduce the visibility of these edges. This smudging, however, will usually produce blurred edges around the object that are inconsistent with the rest of the image.
All objects in an image must also contain the proper perspective and geometry. The perspective of three-dimensional objects in the two-dimensional image is dictated by the viewing geometry and the camera. If the geometry of an object is inconsistent with the other objects in the image, then it was probably added from another image. For example, lines that are parallel in the scene will converge to a point called the vanishing point in the image. If the parallel lines of an object do not converge to the same vanishing point as the rest of the image, then the object could not have been imaged with the same camera or perspective as the rest of the image.
Finally, an understanding of how image processing alters the image characteristics can lead to signs of alteration. For example, when the image contrast is enhanced, the resulting gray-level histogram of the image will usually display "holes" or gray-level values that contain are no longer present in the image. An object from one image that is inserted into a second image may exhibit a different histogram that will indicate that it was not originally part of the second image. However, if an image has been enhanced using an adaptive processing algorithm, then the image characteristics, such as the gray-level histogram or the edge sharpness, can change locally even though no other alteration have been made. Adaptive processing should not be used on real images if the integrity of the image is to be preserved.
Note: If you have any image to check its authentity, please send us the same and we will do the needful.