Procedia Computer Science · 2020 · 60 citations · 11 references
ComputingComputer ScienceComputer VisionDigital ForensicsDigital ImagingComputer ForensicsDigital Image ProcessingDigital Signal ProcessingForgery DetectionImage AnalysisImage ProcessingCopy-move Forgery DetectionDiscrete Fourier TransformImage ForensicsImage Forgery DetectionImage SplicingSplicing DetectionStatistical Image FeaturesSupport Vector Machine Classification
Majority of the existing detection algorithms are able to deal with either type of forgery (splicing or copy-move). However, if we require a unified approach, we need to combine two previous works for each one of the alterations. In order to solve this problem, the authors present a new algorithm for the detection of splicing and copy-move forgery in the same instance. In this paper, a forgery detection technique is proposed which exploits the artifacts originated due to manipulations performed on JPEG encoded images. In JPEG compression technique, an image is divided into non-overlapping blocks of size 8x8 pixels and discrete cosine transform (DCT) coefficients are evaluated for each block independently. When a JPEG compressed image is tampered, there is a change in the statistical properties of AC components of block DCT coefficients. To capture this change, we propose to use standard deviation and count of non-zero DCT coefficients corresponding to each of the AC frequency components independently. The images are cropped by removing a few rows and columns from the top left corner and suggested features are evaluated for test image and its cropped version. The extracted feature vector is used with the support vector machine (SVM) for the classification of authentic and forged images. Experiments are conducted on a standard dataset of pre- and post-processed forged images CASIA vl.O and v2.0 to consolidate the theoretical concept of the proposed technique. Also, the comparative analysis is performed to showcase better detection rates compared with the state-of-the-art methods.
11
CASIA Image Tampering Detection Evaluation Database
Jing Dong, Wei Wang, Tieniu Tan · 2013 · 767 citations
Digital image splicing detection based on Markov features in DCT and DWT domain
Zhongwei He, Wei Lu, Wei Sun et al. · Pattern Recognition · 2012 · 313 citations
A natural image model approach to splicing detection
Yun Q. Shi, Chunhua Chen, Wen Chen · 2007 · 226 citations