Image forgery detection based on statistical features of block DCT coefficients

Dua, Shilpa; Singh, Jyotsna; Parthasarathy, Harish

doi:10.1016/j.procs.2020.04.038

Cited by 46 publications

(20 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The weight of each DCT block is determined through the Euclidean distance of the Gaussian model. And the r th feature of the feature map is processed through the following equation [ 14 ]:

…”

Section: Improved Generative Adversarial Network By Ccamentioning

confidence: 99%

Optimization of a Deep Learning Algorithm for Security Protection of Big Data from Video Images

Geng¹,

Yan

Lu³

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

With the rapid development of communication technology, digital technology has been widely used in all walks of life. Nevertheless, with the wide dissemination of digital information, there are many security problems. Aiming at preventing privacy disclosure and ensuring the safe storage and sharing of image and video data in the cloud platform, the present work proposes an encryption algorithm against neural cryptography based on deep learning. Primarily, the image saliency detection algorithm is used to identify the significant target of the video image. According to the significant target, the important region and nonimportant region are divided adaptively, and the encrypted two regions are reorganized to obtain the final encrypted image. Then, after demonstrating how attackers conduct attacks to the network under the ciphertext attack mode, an improved encryption algorithm based on selective ciphertext attack is proposed to improve the existing encryption algorithm of the neural network. Besides, a secure encryption algorithm is obtained through detailed analysis and comparison of the security ability of the algorithm. The experimental results show that Bob’s decryption error rate will decrease over time. The average classification error rate of Eve increases over time, but when Bob and Alice learn a secure encryption network structure, Eve’s classification accuracy is not superior to random prediction. Chosen ciphertext attack-advantageous neural cryptography (CCA-ANC) has an encryption time of 14s and an average speed of 69mb/s, which has obvious advantages over other encryption algorithms. The self-learning secure encryption algorithm proposed here significantly improves the security of the password and ensures data security in the video image.

show abstract

“…The weight of each DCT block is determined through the Euclidean distance of the Gaussian model. And the r th feature of the feature map is processed through the following equation [ 14 ]:

…”

Section: Improved Generative Adversarial Network By Ccamentioning

confidence: 99%

Optimization of a Deep Learning Algorithm for Security Protection of Big Data from Video Images

Geng¹,

Yan

Lu³

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

show abstract

“…Block-based techniques separate the image into overlapping or non-overlapping squares or blocks of circle shapes as shown in Figure 5. Then using an effective feature transform, features can be extracted from each block such as discrete cosine transform [37,38], discrete wavelet transform [39], curvelet transform [40], fourier transform [41,42,43], fast walsh-hadamard transform (fwht), singular value decomposition [44], principal component analysis, intensity, zernike moments [45], and combinations of them. Some of multiscale decomposition transform (MSD) like pyramid and wavelet transform lacks directionality.…”

Section: Ic-mfd Algorithms Based Transform Domainmentioning

confidence: 99%

A comparative analysis of image copy-move forgery detection algorithms based on hand and machine-crafted features

Ahmed

Hammad

Jamil

2021

IJEECS

View full text Add to dashboard Cite

<span>Digital image forgery (DIF) is the act of deliberate alteration of an image to change the details transmitted by it. The manipulation may either add, delete or alter any of the image features or contents, without leaving any hint of the change induced. In general, copy-move forgery, also referred to as replication, is the most common of the various kinds of passive image forgery techniques. In the copy-move forgery, the basic process is copy/paste from one area to another in the same image. Over the past few decades various image copy-move forgery detection (IC-MFDs) surveys have been existed. However, these surveys are not covered for both IC-MFD algorithms based hand-crafted features and IC-MFDs algorithms based machine-crafted features. Therefore, The paper presented a comparative analysis of IC-MFDs by collect various types of IC-MFDs and group them rely on their features used. Two groups, i.e. IC-MFDs based hand-crafted features and IC-MFDs based machine-crafted features. IC-MFD algorithms based hand-crafted features are the algorithms that detect the faked image depending on manual feature extraction while IC-MFD algorithms based machine-crafted features are the algorithms that detect the faked image automatically from image. Our hope that this presented analysis will to keep up-to-date the researchers in the field of IC-MFD.</span>

show abstract

“…Methods Analysed in this Study: In our study, we have chosen five methods [24], [25], [26], [27], and [28], that are based on statistical approaches and use hand-crafted features and white-box models (classical machine learning algorithms) to classify pristine and tampered images. Alahmadi et al [24] used textural pattern LBP and frequency domain DCT to extract features for classification using SVM.…”

Section: Related Workmentioning

confidence: 99%

“…The method achieved 97.00% on CASIA v1.0, 97.50% on CASIA v2.0 and 97.77% on the Columbia Colored dataset. Shilpa et al [25] made use of artifacts that originated due to manipulations of JPEG encoded images, extracting useful features through standard deviation and number of ones in the AC components of DCT coefficients. The achieved average detection rates for CASIA v1.0 and CASIA v2.0 are above 93% and 98% respectively.…”

Section: Related Workmentioning

confidence: 99%

“…IV. OVERVIEW OF THE ANALYSIS Out of our selected papers [24], [25], [26], [27], and [28] all have either explicitly mentioned that they are solving for both copy-move and splice forgeries or they have used a dataset that has both copy-move and spliced images, thus allowing us to assume that they have considered both type of forgeries.…”

Section: Datasets Descriptionmentioning

confidence: 99%

See 1 more Smart Citation

Analysing Statistical methods for Automatic Detection of Image Forgery

Masud¹,

Agarwal²

2021

Preprint

View full text Add to dashboard Cite

Image manipulation and forgery detection have been a topic of research for more than a decade now. New-age tools and large-scale social platforms have given space for manipulated media to thrive. These media can be potentially dangerous and thus innumerable methods have been designed and tested to prove their robustness in detecting forgery. However, the results reported by state-of-the-art systems indicate that supervised approaches achieve almost perfect performance but only with particular datasets. In this work, we analyze the issue of out-ofdistribution generalisability of the current state-of-the-art image forgery detection techniques through several experiments. Our study focuses on models that utilise handcrafted features for image forgery detection. We show that the developed methods fail to perform well on cross-dataset evaluations and in-the-wild manipulated media. As a consequence, a question is raised about the current evaluation and overestimated performance of the systems under consideration.Note: This work was done during summer research internship at ITMR Lab, IIIT-Allahabad under supervision of Prof. Anupam Agarwal.

show abstract

Image forgery detection based on statistical features of block DCT coefficients

Cited by 46 publications

References 12 publications

Optimization of a Deep Learning Algorithm for Security Protection of Big Data from Video Images

Optimization of a Deep Learning Algorithm for Security Protection of Big Data from Video Images

A comparative analysis of image copy-move forgery detection algorithms based on hand and machine-crafted features

Analysing Statistical methods for Automatic Detection of Image Forgery

Contact Info

Product

Resources

About