Image Processing Reference
In-Depth Information
methods such as JPEG compression or the addition of noise can reduce the accuracy
of the detection algorithm.
6.2
Copy-Move Forgery (CMF)
One of the most common used digital image forgery methods is a copy-move forgery
(CMF) [20], where part of an image is copied and moved to another location in the
same image. The purpose of this kind of forgery is usually to hide or to add some
content or object in the image. Since the forged region came from the same image,
it is impossible to use some statistical properties (for example, camera noise or
illumination conditions) for forgery detection because they are very well matched
within the image. Taking the forged region from the same image also simplifies the
forgery process because it is easier to fit the forged region into the image due to the
similarity of properties of the copied region and the rest of image.
Plain copy-move forgery is a type of forgery where the copied area is translated
to a new location in the same image without changing any properties of the copied
area. Therefore, in that kind of forgery, there are two completely identical areas in
the image which makes plain copy-move forgery detection quite easy to implement.
More complicated types of forgery can be done by transformation of a copied
region before translation to a new location. Some possible transformations of copied
regions are:
1. scaling - enlarging or shrinking of a copied area by an equal scale factor in all
directions,
2. rotation - circular moving of a copied area around a centre of rotation by a arbi-
trary angle,
3. distortion - enlarging or shrinking of a copied area by a scale factor that is not
the same in all directions,
4. combination - application of more than one transformation of a copied area.
The result of applying any of these transformations is a change in the proper-
ties of the copied area. Therefore, searching for forgeries is not as simple as in the
case of plain CMF. Figure 6.1 shows some examples of CMFs from the CoMoFoD
database [24].
Hiding of forgery traces can be done by applying some post-processing methods.
It is possible to apply a post-processing method on the whole image after forgery,
but sometimes post-processing is applied only on copied region borders to assure
better fitting with the new background. The most common post-processing methods
used in digital image forgery are JPEG compression, addition of noise and image
blurring (Fig. 6.2).
6.3
Copy-Move Forgery Detection (CMFD)
Detection of copy-move forgery has been widely researched [9]. Developed meth-
ods for copy-move forgery detection can be categorized as keypoint-based and
Search WWH ::
Custom Search