Exposing Video Forgery Detection Using Intrinsic ...

76 downloads 459911 Views 674KB Size Report
software's like Adobe Photoshop, Corel Draw or Gimp are available for the modification of the ... machine-learning based and requires minimal user interaction.
International Journal of Recent Advances in Engineering & Technology (IJRAET) _______________________________________________________________________________________________

Exposing Video Forgery Detection Using Intrinsic Fingerprint Traces 1 1,2,3

Saranya.R , 2Saranya. S, 3Cristin.R

Dept of Information Technology, St. Joseph College of Engineering

Abstract- One of the principal problems in video forensics is determining if a particular video is authentic or not. This can be a crucial task when video’s are used as basic evidence to influene judgement like, for example, in a court of law. We proposed to address passive forgeries detection in a digital video based on the statistical property of noise residue. To analyse the temporal correlation of block-level noise residue to locate the tampered region of a video. Extensive simulation results are presented to confirm that the technique is able to precisely individuate the tampered region and in addition, to estimate transformation parameters with high reliability. Index Terms – Bayesian, Grayscale, superpixel, Weiner filter

I . INTRODUCTION In recently years, due to the advances of network technologies, low-cost multimedia devices, sophisticated image/video editing software and wide adoptions of digital multimedia coding standards, digital multimedia applications have become increasingly popular in our daily life. However, the digital nature of the media files, they can now be easily manipulated, synthesized and tampered in numerous ways without leaving visible clues. As a result, the integrity of image/video content can no longer be taken for granted and a number of forensic-related issues arise. Furthermore in media outlets, scientific journals, political campaigns, courtrooms, and photo hoaxes that land in our email boxes, forged images are appearing more frequently in a unique way unable to identify the fake image with the needed sophistication. Authenticity and integrity of the digital images are well-thought-out to be important to overcome these issues because of the forging in fields such as forensic, medical imaging, e-commerce, industrial photography, etc. The authenticity verification check of the image is popularly used where the images are considered to supporting evidences, historical records, insurance claims, etc. Conversely, with encouragements of the today’s computer technology, more sophisticated software’s like Adobe Photoshop, Corel Draw or Gimp are available for the modification of the original images, resulting in image tampering. This project uses an artificial neural network to store the extracted data from image streams to completely and accurately detect tampered locations particularly.We propose a forgery

detection method that exploits inconsistencies in the pixelof the illumination of images. Our approach is machine-learning based and requires minimal user interaction. The technique is applicable to images containing two or more people and requires no expert interaction for the tampering decision.

II. LITERATURE SURVEY Over the past decade, many efforts have been made in passive image forensics. Although it is able to detect tampered images at high accuracies based on some carefully designed mechanisms, localization of the tampered regions in a fake image still presents many challenges, especially when the type of tampering operation is unknown. Some researchers have realized that it is necessary to integrate different forensic approaches in order to obtain better localization performance. In this paper, we propose a framework to improve the performance of forgery localization via integrating tampering possibility maps. In the proposed framework, we first select and improve two existing forensic approaches, i.e., statistical feature based detector and copy-move forgery detector, and then adjust their results to obtain tampering possibility maps[1]. In this paper, we analyze one of the most common forms of photographic manipulation, known as image composition or splicing. We propose a forgery detection method that exploits subtle inconsistencies in the color of the illumination of images. Our approach is machine-learning based and requires minimal user interaction. The technique is applicable to images containing two or more people and requires no expert interaction for the tampering decision. To achieve this, we incorporate information from physics- and statistical-based illuminant estimators on image regions of similar material [2]. We describe a geometric technique to detect physically implausible trajectories of objects in video sequences. This technique explicitly models the three-dimensional ballistic motion of objects in free-flight and the two-dimensional projection of the trajectory into the image plane of a static or moving camera. Deviations from this model provide evidence of manipulation [3]. In this paper, the problem of detecting if an image has been forged is investigated; in particular, attention has been paid to the case in which an area of an image is

_______________________________________________________________________________________________ ISSN (Online): 2347 - 2812, Volume-5, Issue -4,5, 2017 73

International Journal of Recent Advances in Engineering & Technology (IJRAET) _______________________________________________________________________________________________ copied and then pasted onto another zone to create a duplication or to cancel something that was awkward. Generally, to adapt the image patch to the new context a geometric transformation is needed. To detect such modifications, a novel methodology based on scale invariant features transform (SIFT) is proposed [4]. In this paper, we analyze the interaction between a forger and a forensic investigator by examining the problem of authenticating digital videos. Specifically, we study the problem of adding or deleting a sequence of frames from a digital video. We begin by developing a theoretical model of the forensically detectable fingerprints that frame deletion or addition leaves behind, then use this model to improve upon the video frame deletion or addition detection technique proposed by Wang and Farid [5].

into something that is more meaningful and easier to analyse. In the segmentation based forgery detection, many of the segmentation technique are adopted. They are Gray scale segmentation Super pixel segmentation Morphological analysis Gray scale segmentation The gray scale segmentation technique is applied to the gray scale image for extracting the desired region for the precise dominion. In grayscale segmentation, primarily the image is converted into gray scale and edge detection function is applied to generate the gradient image based on which the image is segmented.

III. DESCRIPTION Here in the proposed system, we focus on tampered region localization for image forensics. A passive image tampering detection method, by an algorithmic process which can locate tampered region(s) in a video based on the statistical properties of noise residue. Passive approach based detection technique are applied for the image forgery detection since it is more advantageous than the active approach which utilize prior knowledge about the image for the authentication. The passive approach acceptance with computational risk is considered tolerable because of its capability to detect the fake images forged with any image forgery. Also this project uses intrinsic fingerprint traces which uses pixel value mapping operation to find the tampered region. This project uses an artificial neural network to store the extracted data from image streams to completely and accurately detect tampered locations particularly. The experimental results shows that the propsed method achieves promising detection accuracy for fine-qualityvideos.

Super pixel segmentation Super pixel segmentation groups the image pixels based on homogeneity criteria and restore the object boundary separating the region. The boundary representation is irregular and it signifies super pixel of the different size. Morphological analysis Morphological based segmentation is a collection of non-linear operations related to the shape or morphology of features in an image. morphological operations rely only on the relative ordering of pixel values, not on their numerical values, and therefore are especially suited to the processing of binary images. B.

Correlation computation

When a region is forged, the correlation value of temporal noise residue in the region is usually changed (increased or decreased) depending on the forgery scheme used. The two inpainting schemes are used in this project to simulate two typical kinds of tampering processes for evaluating the performance of the proposed forgery detection schemes. C.

Noise residue

The noise residual can be easily extracted by subtracting the original image from its noise-free version. The noise residue consists PNU and highspatial-frequency details of image content. The PNU was extracted from noise residual via an averaging operation. D.

Algorithm INPUT: Ѕ , λ , Т , К INITIALIZE: Choose w1 s.t. ‖w1‖ ≤1/√λ FOR t = 1,2,….,Т

Figure 1: Overview of propsed Architecture A.

Image segmentation

Choose At ⊆ Ѕ, where | At |=К +

Set At = {(x,y) ε At : 𝓎 wt , x < } Segmentation is the process of partitioning a digital 1 image into multiple segments (sets of pixels, also Set ηt = λt known as super pixels). The goal of segmentation is to simplify and/or change the representation of an image _______________________________________________________________________________________________ ISSN (Online): 2347 - 2812, Volume-5, Issue -4,5, 2017 74

International Journal of Recent Advances in Engineering & Technology (IJRAET) _______________________________________________________________________________________________ Set wt+1 = (1 - ηt λ)wt + 2

Set wt+1 = min { 1,

ηt k

1 √λ ∥w

1∥ t +2

+

(x,y)ϵA t

+𝓎x

} wt+1 2

OUTPUT: WT+1 The algorithm, called Pegasos, is a modified stochastic gradient method in which every gradient descent step is accompanied with a projection step. At each iteration, a single training example is chosen at random and used to estimate a sub-gradient of the objective, and a step with pre-determined step-size is taken in the opposite direction.

Figure 3: Train the input image

n

Given a training set S = { (xi,yi) } mi=1, where xi ϵ R & yi ϵ {+1,-1} , in order to find the minimizer of the problem min λ w2

∥ w ∥2 +

1 m

x,y ϵs l(w;

x, y )

(1)

Where, l w; x, y

= max⁡{0,1 − y < w, x >}

(2)

The algorithm receives as input two parameters: T - the number of iterations to perform; k - the number of examples to use for calculating sub-gradients. Initially, we set w1 to any vector whose norm is at most 1/√λ. On iteration t of the algorithm, choose a set At ⊆ S of size k Figure 4:Segmented Image

IV. EXPERIMENTAL RESULTS Thus this system achieves promising accuracy ,as it can easily extended to work with traces left by other kind of processing.

Figure 5:Identification of forged region

V.CONCLUSION

Figure 2: Test Image

In this paper, we have proposed a digital video forgery detection scheme using temporal noise correlation without the need of embedding any prior digital signature in the compressed video. We have also proposed a statistical classification scheme. Consequently, a Bayesian classifier is used to find the optimal threshold value based on the estimated parameters. Experimental results show that the proposed method achieves promising detection accuracy for finequality videos.

REFERENCES [1].

HaodongLi, WeiqiLuo, Xiaoqing Qiu (2017), ”Image Forgery Localization via Integrating Tampering Possibility Maps”, in IEEE

_______________________________________________________________________________________________ ISSN (Online): 2347 - 2812, Volume-5, Issue -4,5, 2017 75

International Journal of Recent Advances in Engineering & Technology (IJRAET) _______________________________________________________________________________________________ Transactions on Information Forensics and Security PP(99):1-1 [2].

[3].

[4].

Tiago José de Carvalho, Christian Riess, Elli Angelopoulou(2013),” Exposing Digital Image Forgeries by Illumination Color Classification”, IEEE Transactions on Information Forensics And Security,Vol.8,no.7. Valentina Conotter, James F. O’Brien, and Hany Farid(2012),” Exposing Digital Forgeries in Ballistic Motion”, IEEE Transaction on Information Forensics And Security,Vol.7,no.1.

[5].

Matthew C. Stamm, W. Sabrina Lin, and K. J. Ray Liu(2012),” Temporal Forensics and AntiForensics for Motion Compensated Video”, in IEEE Transaction on Information Forensics And Security,Vol.7,no.4.

[6].

Tiziano Bianchi, and Alessandro Piva(2012),” Image Forgery Localization via Block-Grained Analysis of JPEG Artifacts”, in IEEE Transactions on Information Forensics and Security,vol.7

[7].

Matthias Kirchner, Peter Winkler and Hany Farid(2013),” Impeding Forgers at Photo Inception”, in SPIE 8665, Media Watermarking, Security, and Forensics.

I Amerini, L Ballan, R Caldelli(2011),”A SIFTBased Forensic Method for Copy–Move Attack Detection and Transformation Recovery”, IEEE Transaction on Information Forensics and Security,Vol.6 

_______________________________________________________________________________________________ ISSN (Online): 2347 - 2812, Volume-5, Issue -4,5, 2017 76