Detecting Hidden Messages Using Higher-Order Statistics and Support Vector Machines

By Siwei Lyu, Hany Farid. Information Hiding, 2002

567 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/637b0ba19f72f9eba326f07c8d46a6e26237e3be

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Low-Complexity Features for JPEG Steganalysis Using Undecimated DCT (2015, methodology): “Selected examples of popular feature sets proposed for detection of steganography in JPEG images are the historically first image quality metric features [1], firstorder statistics of wavelet coefficients [8], Markov features formed by sample intra-block conditional probabilities [29], inter- and in”
Feature-Based Steganalysis for JPEG Images and Its Implications for Future Design of Steganographic Schemes (2004, background): “Based on the previous work of Eggers [8], Tzschoppe [9] constructed a JPEG steganographic scheme (HPDM) that is undetectable using Farid’s scheme.”
Benchmarking for Steganography (2008, background): “There are basically two choices: 1) analytical models, in which c is modeled as a sequence of iid random variables [28] (or Markov chains [39]) and 2) high-dimensional models based on features extracted from the cover/stego objects [32, 13, 2, 1, 43, 36, 26, 11].”
Deep Residual Network for Steganalysis of Digital Images (2019, background): “This trend has been started by Avcibas et al. [1], [2] and Farid and Siwei [18] in early 2000’s and was later greatly improved by representing images with higher-order statistics of noise residuals or DCT coefficients [43], [49], [68].”
Digital image forensics: a booklet for beginners (2010, background): “In a later work [7], the same authors improve their approach by designing new features based on image quality metrics [3] and higher-order statistics of the image [67].”
Steganalysis by Subtractive Pixel Adjacency Matrix (2009, methodology): “The idea of using filtering to enhance signal to noise ratio in steganalysis has been already used, for example, in the WAM features calculating moments from noise residual in Wavelet domain and it implicitly appeared in the construction of Farid’s features [6] and in [40].”
How realistic is photorealistic? (2005, methodology): “In order to capture some of these higher order statistical correlations , we collect a second set of statistics that are based on the errors in a linear predictor of coefficient magnitude [2].”
Rich Models for Steganalysis of Digital Images (2012, background): “4 in [22]) and then perfected through a long series of papers examples of which are [2], [1], [9], [18], [33], [38].”
Using decision fusion of feature selection in digital forensics for camera source model identification (2012, background): “By applying wavelet features for feature optimization, the results of our experiments indicate consistent performance with [24,27,40].”
Blind source camera identification (2004, methodology): “• Wavelet domain statistics Inspired by Farid’s work [3], we decomposed each color band of the image using separable quadratic mirror filters and then calculated the mean for each of the 3 resulting sub-bands (9 features).”
Deep Learning Hierarchical Representations for Image Steganalysis (2017, methodology): “In practice, many steganalysis schemes [17], [19] implement the predictor by convolving a finite impulse response filter K ′ with image X :”
Digital camera identification from sensor pattern noise (2006, methodology): “In their method, each image is represented using a vector of numerical features extracted from it.”
Steganalysis using higher-order image statistics (2006, methodology): “Specifically, in [18], we showed how a statistical model based on first- and higher-order magnitude statistics extracted from a wavelet decomposition, coupled with a linear discriminant analysis (LDA), could be used to detect steganography in grayscale images.”