S. Geetha

Decision tree based light weight intrusion detection using a wrapper approach

The objective of this paper is to construct a lightweight Intrusion Detection System (IDS) aimed at detecting anomalies in networks. The crucial part of building lightweight IDS depends on preprocessing of network data, identifying important features and in the design of efficient learning algorithm that classify normal and anomalous patterns. Therefore in this work, the design of IDS is investigated from these three perspectives. The goals of this paper are (i) removing redundant instances that causes the learning algorithm to be unbiased (ii) identifying suitable subset of features by employing a wrapper based feature selection algorithm (iii) realizing proposed IDS with neurotree to achieve better detection accuracy. The lightweight IDS has been developed by using a wrapper based feature selection algorithm that maximizes the specificity and sensitivity of the IDS as well as by employing a neural ensemble decision tree iterative procedure to evolve optimal features. An extensive experimental evaluation of the proposed approach with a family of six decision tree classifiers namely Decision Stump, C4.5, Naive Bayes Tree, Random Forest, Random Tree and Representative Tree model to perform the detection of anomalous network pattern has been introduced.

Varying radix numeral system based adaptive image steganography

A novel image steganographic scheme based on the primitive and renowned numerical model is proposed. In this adaptive scheme the data to be embedded is dissected into numerals, each having variable information carrying capacity. The dissection is based on the statistics of the host image pixels i.e. the amount of adulteration that a pixel can tolerate. The proposed method provides proficient visual quality despite high payload capacity. The experimental results evaluated on 300 natural images show that the new scheme is resistant to RS steganalysis and offers high visual quality when compared with typical LSB-based schemes as well as their edge adaptive versions, like pixel-value-differencing-based schemes.

Evolving GA Classifier for Breaking the Steganographic Utilities: Stools, Steganos and Jsteg

Differentiating anomalous image documents (stego image) from pure image file (cover image) is difficult and tedious. Steganalytic techniques strive to detect whether an image contains a hidden message or not. This paper presents a genetic algorithm (GA) based approach to image steganalysis. The basic idea is that, the various image quality metrics calculated on cover image files and on stego-image files vis-a-vis their denoised versions, are statistically different. GA is employed to derive a set of classification rules from image data using these image quality metrics and the support-confidence framework is utilized as fitness function to judge the quality of each rule. The generated rules are then used to detect or classify the image documents in a real-time environment. Unlike most existing GA-based approaches, because of the simple representation of rules and the effective fitness function, the proposed method is easier to implement while providing the flexibility to generally detect any new steganography technique. The implementation of the GA based image steganalyzer relies on the choice of these image quality metrics and the construction of a two-class classifier, which will discriminate between the adulterated and the untouched image samples. Experimental results show that the proposed technique provides promising detection rates.

StegoBreaker: Audio Steganalysis using Ensemble Autonomous Multi-Agent and Genetic Algorithm

The goal of steganography is to avoid drawing suspicion to the transmission of a hidden message in multi-medium. This creates a potential problem when this technology is misused for planning criminal activities. Differentiating anomalous audio document (stego audio) from pure audio document (cover audio) is difficult and tedious. Steganalytic techniques strive to detect whether an audio contains a hidden message or not. This paper investigates the use of genetic algorithm (GA) to aid autonomous intelligent software agents capable of detecting any hidden information in audio files, automatically. This agent would make up the detection agent in an architecture comprising of several different agents that collaborate together to detect the hidden information. The basic idea is that, the various audio quality metrics (AQMs) calculated on cover audio signals and on stego-audio signals vis-a-vis their denoised versions, are statistically different. GA employs these AQMs to steganalyse the audio data. The overall agent architecture will operate as an automatic target detection (ATD) system. The architecture of ATD system is presented in this paper and it is shown how the detection agent fits into the overall system. The design of ATD based audio steganalyzer relies on the choice of these audio quality measures and the construction of a GA based rule generator, which spawns a set of rules that discriminates between the adulterated and the untouched audio samples. Experimental results show that the proposed technique provides promising detection rates

Evolving optimised decision rules for intrusion detection using particle swarm paradigm

The aim of this article is to construct a practical intrusion detection system (IDS) that properly analyses the statistics of network traffic pattern and classify them as normal or anomalous class. The objective of this article is to prove that the choice of effective network traffic features and a proficient machine-learning paradigm enhances the detection accuracy of IDS. In this article, a rule-based approach with a family of six decision tree classifiers, namely Decision Stump, C4.5, Naive Bayes Tree, Random Forest, Random Tree and Representative Tree model to perform the detection of anomalous network pattern is introduced. In particular, the proposed swarm optimisation-based approach selects instances that compose training set and optimised decision tree operate over this trained set producing classification rules with improved coverage, classification capability and generalisation ability. Experiment with the Knowledge Discovery and Data mining (KDD) data set which have information on traffic pattern, during normal and intrusive behaviour shows that the proposed algorithm produces optimised decision rules and outperforms other machine-learning algorithm.

Blind image steganalysis based on content independent statistical measures maximizing the specificity and sensitivity of the system

This paper reports the design principles and evaluation results of a new experimental universal, blind image steganalysing system. This system investigates the use of content independent statistical evidences left by the steganograms, as features for an image steganalyzer. The work is aimed at maximizing the sensitivity and specificity of the steganalyzer and to accomplish both security and system performance. A genetic-X-means classifier is constructed to realize the proposed model. For performance evaluation, a database composed of 5600 plain and stego images (generated by using seven different embedding schemes) was established. The results of our empirical experiment prove the vitality of the proposed scheme in detecting stego anomalies in images. In addition, the simulation results show that the effectiveness of steganalytic system can be enhanced by considering the content independent distortion measures and maximizing the sensitivity and specificity of the system.

Audio Steganalysis using Ensemble of Autonomous Multi-Agent and Support Vector Machine Paradigm

This paper investigates the use of support vector machines (SVM) to create and train agents capable of detecting any hidden information in audio files. This agent would make up the detection agent in an architecture comprising of several different agents that collaborate together to detect the hidden information. The system exploits a soft computing approach to detect the presence of hidden messages in audio signals, by using the audio quality metrics. The distribution of various statistical distance measures, calculated on cover audio signals and on stego-audio signals vis-a-vis their denoised versions, are different. The overall agent architecture operates as an automatic target detection (ATD) system. The architecture of ATD system is presented in this paper and it is shown how the detection agent fits into the overall system. The design of ATD based audio steganalyzer relies on the choice of these audio quality measures and the construction of a SVM classifier, which discriminates between the adulterated and the untouched audio samples

StegoHunter: Steganalysis of LSB Embedded Images Based on Stego-Sensitive Threshold Close Color Pair Signature

This paper proposes a reliable framework for the detection of the least significant bit (LSB) steganography using digital media files as cover objects. Steganographic methods attempt to insert data in multimedia signals in an undetectable fashion. However, these methods often disrupt the underlying signal characteristics, thereby allowing detection under careful steganalysis. Under repeated embedding, disruption of the signal characteristics is the highest for the first embedding and decreases subsequently. This principle is used to derive a steganalysis tool that detects the presence of hidden messages in uncompressed twenty-four bits BMP image. This work presents close color pair analysis with stego-sensitive threshold (CCPASST) to detect stego-objects with even 10% payload. In earlier works 20% payload was detected through close color pair analysis. The new framework exploits the first-order statistics of structural similarity index measure of the samples to calculate the threshold. The literature contains only one other detector specialized with variable threshold, and the one presented here is substantially more sensitive. Simulation results with the stego-sensitive threshold applied on well-known LSB steganographic technique indicate that this approach is superior to the earlier methods and is able with promising accuracy to distinguish between clean and stego images.

An Active Rule Based Approach to Audio Steganalysis with a Genetic Algorithm

Differentiating anomalous audio document (stego audio) from pure audio document (cover audio) is difficult and tedious. Steganalytic techniques strive to detect whether an audio contains a hidden message or not. This paper presents a genetic algorithm (GA) based approach to audio steganalysis. The basic idea is that, the various audio quality metrics calculated on cover audio signals and on stego audio signals vis-avis their denoised versions, are statistically different. GA is employed to derive a set of classification rules from audio data using these audio quality metrics, and the support-confidence framework is utilized as fitness function to judge the quality of each rule. The generated rules are then used to detect or classify the audio documents in a real-time environment. Unlike most existing GA-based approaches, because of the simple representation of rules and the effective fitness function, the proposed method provides flexibility to generally detect any new steganography technique. The implementation of the GA based audio steganalyzer relies on the choice of these audio quality metrics and the construction of a two-class classifier, which will discriminate between the adulterated and the untouched audio samples. Experimental results show that the proposed technique provides promising detection rates.

Geometric Attack Invariant Watermarking with Biometric Data - Applied on Offline Handwritten Signature

Biometric watermarking refers to the process of incorporating the handwritten signatures or fingerprints in watermarking technology. In this paper, we present a novel statistical attack invariant watermarking scheme to embed an offline handwritten signature invisibly in the host as a notice of genuine ownership. The scheme employs Discrete Cosine Transform (DCT) and Singular Value Decomposition (SVD) with Least Significant Bit (LSB) encoding for watermark insertion, which is to be known as DCT-SVD scheme. Experimental results confirm that DCT-SVD scheme is robust to statistical attacks even in the presence of deliberate distortions.