The first block based detection technique
was introduced by Fridrich et al. 12, using a block
matching detection scheme based on discrete cosine transform (DCT). Several
other detection techniques based on DCT are proposed to improve the accuracy,
computational complexity and robustness against post-processing operations in 14, 7, 43. Principal component
analysis (PCA) feature is employed to detect the forgery by Popescu and Farid 25. A new technique
was introduced by Luo et al. 22, in which
blocks are divided further into four sub-blocks. Authors claim that this technique
is robust against post-processing operations such as JPEG compression, Gaussian
blurring, and additive noise. Pan et al. 24, developed a technique
based on scale invariant feature transform (SIFT), which is robust to distortions
of the duplicated regions. Bashar et al. 4, described a new
block based approach to detect copy-move forgery using discrete wavelet
transform (DWT) and kernel principal component analysis (KPCA). This technique
is robust against the JPEG compression and additive noise. Silva et al.31, developed a technique
based on multi-scale analysis and voting processes of a digital image by
employing Speeded-Up Robust Features (SURF) technique. Lee 16, introduced a technique
using Gabor filter. Ardizzone et al. 3 developed an
approach based on matching triangles. Cozzolino et al. 10 proposed a technique
based on  dense-field techniques and
Zernike moments. Li et al. introduced a detection technique 18, based on  vlFeat software and RANSAC.

Pun et al. 26 described a
detection technique by merging keypoint features and block based features. A new
technique based on histogram of orientated gradients was presented  by Lee et al. 17. Gürbüz et al. 13, introduced a new
approach using circular projection. By employing split-half recursion matching
strategy with SIFT keypoints Zhao et al. 42 introduced an
algorithm. Particle swam optimization (PSO) along with SIFT keypoint is taken
as the basis for detecting forgery, Wenchang et al. proposed a new technique 39 but his technique
is unable to detect forgery when duplicated region are very small. Zandi et al.
proposed a technique 41 based on
interest point detector. Ferreira et al. in 
11 proposed a technique
based on Behavior knowledge space-based fusion. Zhu et al. in 44 introduced a technique
that employs scaled ORB along with RANSAC algorithm. A Multi-Level Dense
Descriptor (MLDD) based detection algorithm was introduced by Bi et al. in 6. Superpixel
content based adaptive feature point detector is employed to detect copy-move
forgery by  Wang et al. in  38. Tralic et al.
in 33 combined cellular
automata(CA) and local binary patterns (LBP) to detect forgery. Recently, Yang
et al.  40, devised an
approach using CMFD-SIFT. This technique improves the invariance to mirror transformation. 

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now