Biblio

Common Vulnerabilities and Exposures (CVE) databases contain information about vulnerabilities of software products and source code. If individual elements of CVE descriptions can be extracted and structured, then the data can be used to search and analyze CVE descriptions. Herein we propose a method to label each element in CVE descriptions by applying Named Entity Recognition (NER). For NER, we used BERT, a transformer-based natural language processing model. Using NER with machine learning can label information from CVE descriptions even if there are some distortions in the data. An experiment involving manually prepared label information for 1000 CVE descriptions shows that the labeling accuracy of the proposed method is about 0.81 for precision and about 0.89 for recall. In addition, we devise a way to train the data by dividing it into labels. Our proposed method can be used to label each element automatically from CVE descriptions.

The success of deep learning based steganography has shifted focus of researchers from traditional steganography approaches to deep learning based steganography. Various deep steganographic models have been developed for improved security, capacity and invisibility. In this work a multi-data deep learning steganography model has been developed using a well known deep learning model called Generative Adversarial Networks (GAN) more specifically using deep convolutional Generative Adversarial Networks (DCGAN). The model is capable of hiding two different messages, meant for two different receivers, inside a single cover image. The proposed model consists of four networks namely Generator, Steganalyzer Extractor1 and Extractor2 network. The Generator hides two secret messages inside one cover image which are extracted using two different extractors. The Steganalyzer network differentiates between the cover and stego images generated by the generator network. The experiment has been carried out on CelebA dataset. Two commonly used distortion metrics Peak signal-to-Noise ratio (PSNR) and Structural Similarity Index Metric (SSIM) are used for measuring the distortion in the stego image The results of experimentation show that the stego images generated have good imperceptibility and high extraction rates.

Data is a collection of information from the activities of the real world. The file in which such data is stored after transforming into a form that machines can process is generally known as data set. In the real world, many data sets are not complete, and they contain various types of noise. Missing values is of one such kind. Thus, imputing data of these missing values is one of the significant task of data pre-processing. This paper deals with two real time health care data sets namely life expectancy (LE) dataset and chronic kidney disease (CKD) dataset, which are very different in their nature. This paper provides insights on various data imputation techniques to fill missing values by analyzing them. When coming to Data imputation, it is very common to impute the missing values with measure of central tendencies like mean, median, mode Which can represent the central value of distribution but choosing the apt choice is real challenge. In accordance with best of our knowledge this is the first and foremost paper which provides the complete analysis of impact of basic data imputation techniques on various data distributions which can be classified based on the size of data set, number of missing values, type of data (categorical/numerical), etc. This paper compared and analyzed the original data distribution with the data distribution after each imputation in terms of their skewness, outliers and by various descriptive statistic parameters.

We present a ternary source coding scheme in this paper, which is a special class of low density generator matrix (LDGM) codes. We prove that a ternary linear block LDGM code, whose generator matrix is randomly generated with each element independent and identically distributed, is universal for source coding in terms of the symbol-error rate (SER). To circumvent the high-complex maximum likelihood decoding, we introduce a special class of convolutional LDGM codes, called block Markov superposition transmission of repetition (BMST-R) codes, which are iteratively decodable by a sliding window algorithm. Then the presented BMST-R codes are applied to construct a tandem scheme for Gaussian source compression, where a dead-zone quantizer is introduced before the ternary source coding. The main advantages of this scheme are its universality and flexibility. The dead-zone quantizer can choose a proper quantization level according to the distortion requirement, while the LDGM codes can adapt the code rate to approach the entropy of the quantized sequence. Numerical results show that the proposed scheme performs well for ternary sources over a wide range of code rates and that the distortion introduced by quantization dominates provided that the code rate is slightly greater than the discrete entropy.

Minimizing embedding impact model of steganography has good performance for steganalysis detection. By using effective distortion cost function and coding method, steganography under this model becomes the mainstream embedding framework recently. In this paper, to improve the anti-detection performance, a new steganography optimization model by constructing a reference cover is proposed. First, a reference cover is construed by performing a filtering operation on the cover image. Then, by minimizing the residual between the reference cover and the original cover, the optimization function is formulated considering the effect of different modification directions. With correcting the distortion cost of +1 and \_1 modification operations, the stego image obtained by the proposed method is more consistent with the natural image. Finally, by applying the proposed framework to the cost function of the well-known HILL embedding, experimental results show that the anti-detection performance of the proposed method is better than the traditional method.

Using Generative Adversarial Networks (GAN) to translate images is a significant field in computer vision. There are partial distortion, artifacts and detail loss in the images generated by current image translation algorithms. In order to solve this problem, this paper adds attention-based residual neural network to the generator of GAN. Attention-based residual neural network can improve the representation ability of the generator by weighting the channels of the feature map. Experiment results on the Facades dataset show that Attention Residual GAN can translate images with excellent quality.

Sharing relational databases on the Internet creates the need to protect these databases. Its output in substantial losses to the data storing systems because of unauthorized access to information that could lose novelty. The research associations use the research databases to mine new information about the research works of the relational databases that are available for free. It is a great challenge to maintain authenticity because these databases are vulnerable to security issues. Watermarking is a candidate solution that fully protects databases shared with the receiver. The protection of relational database ownership that may continue to evolve against the various aquatic mechanisms shared with the recipient that arouses appetite for attacks and must continue to evolve so that they can have database knowledge to support their decision-making system is effective. The relational database based onVirtual private key Watermarking using numeric attribute) involves embedding the same watermark in the same properties in different places in the same place. Therefore, data attackers cannot remove watermarks from data. The proposed strategy is to work by inserting watermark bits in such a way that it causes minimal distortion in the data and the data usability must remain intact after the data is watermarked. The proposed strategy is to work by inserting watermark bits in such a way that it causes minimal distortion in the data and the ability to use the data after watermarking the data must remain intact. The existence of a primary key is the main feature or compulsory item for most of the strategies. Our method provides solutions no primary key feature where the integrating search system of the database remains intact after watermarking distortion.

In this paper, we propose a zero-distortion blind reversible robust scheme for watermarking relational databases called PKMark. Data owner can declare the copyright of the databases or pursue the infringement by extracting the water-mark information embedded in the database. PKMark is mainly based on the primary key attribute of the tuple. So it does not depend on the type of the attribute, and can provide high-precision numerical attributes. PKMark uses RSA encryption on the watermark before embedding the watermark to ensure the security of the watermark information. Then we use RSA to sign the watermark cipher text so that the owner can verify the ownership of the watermark without disclosing the watermark. The watermark embedding and extraction are based on the hash value of the primary key, so the scheme has blindness and reversibility. In other words, the user can obtain the watermark information or restore the original database without comparing it to the original database. Our scheme also has almost excellent robustness against addition attacks, deletion attacks and alteration attacks. In addition, PKMark is resistant to additive attacks, allowing different users to embed multiple watermarks without interfering with each other, and it can indicate the sequence of watermark embedding so as to indicate the original copyright owner of the database. This watermarking scheme also allows data owners to detect whether the data has been tampered with.

This paper proposes a novel model to detect Deep-Fakes, which are hyper-realistic fake videos generated by advanced AI algorithms involving facial superimposition. With a growing number of DeepFakes involving prominent political figures that hold a lot of social capital, their misuse can lead to drastic repercussions. These videos can not only be used to circulate false information causing harm to reputations of individuals, companies and countries, but also has the potential to cause civil unrest through mass hysteria. Hence it is of utmost importance to detect these DeepFakes and promptly curb their spread. We therefore propose a CNN-based model that learns inherently distinct patterns that change between a DeepFake and a real video. These distinct features include pixel distortion, inconsistencies with facial superimposition, skin colour differences, blurring and other visual artifacts. The proposed model has trained a CNN (Convolutional Neural Network), to effectively distinguish DeepFake videos using a frame-based approach based on aforementioned distinct features. Herein, the proposed work demonstrates the viability of our model in effectively identifying Deepfake faces in a given video source, so as to aid security applications employed by social-media platforms in credibly tackling the ever growing threat of Deepfakes, by effectively gauging the authenticity of videos, so that they may be flagged or ousted before they can cause irreparable harm.

Steganography is the method of hiding one type of information into other type of information, hiding a secret a message in a cover so that others can't know the presence of the secret information. It provides an extra layer of security in communication and information sharing. Security is an important aspect of the communication process; everyone want security in communication. The main purpose of this paper is to introduce security of information that people share among them. In this paper we are presenting different methods of substitution techniques of image steganography and their comparison. Least significant bit and most significant bit substitution techniques are used. Information is hidden in an image file and then decoded back for the secret message. Hiding the presence of any hidden information makes this more secure. This implementation can be used by secret service agencies and also common people for secure communication.

In recent years, the application of style transfer has become more and more widespread. Traditional deep learning-based style transfer networks often have problems such as image distortion, loss of detailed information, partial content disappearance, and transfer errors. The style transfer network based on deep learning that we propose in this article is aimed at dealing with these problems. Our method uses image edge information fusion and semantic segmentation technology to constrain the image structure before and after the migration, so that the converted image maintains structural consistency and integrity. We have verified that this method can successfully suppress image conversion distortion in most scenarios, and can generate good results.

Arbitrary style transfer means that stylized images can be generated from a set of arbitrary input image pairs of content images and style images. Recent arbitrary style transfer algorithms lead to distortion of content or incompletion of style transfer because network need to make a balance between the content structure and style. In this paper, we introduce a dual attention network based on style attention and channel attention, which can flexibly transfer local styles, pay more attention to content structure, keep content structure intact and reduce unnecessary style transfer. Experimental results show that the network can synthesize high quality stylized images while maintaining real-time performance.

The circulating current in the D-winding may cause primary current waveform distortion, and the reliability of the restraint criterion based on the typical magnetizing inrush current characteristics will be affected. The magnetizing inrush current with typical characteristics is the sum of primary current and circulating current. Using the circulating current to compensate the primary current can improve the reliability of the differential protection. When the phase is not saturated, the magnetizing inrush current is about zero. Therefore, the primary current of unsaturated phase can be replaced by the opposite of the circulating current. Based on this, an engineering practical calculation method for circulating current is proposed. In the method, the segmented primary currents are used to replace the circulating current. Phasor analysis is used to demonstrate the application effect of this method when remanence coefficients are different. The method is simple and practical, and has strong applicability and high reliability. Simulation and recorded waveforms have verified the effectiveness of the method.

Robustness, both to accident and to malevolent perturbations, is a crucial determinant of the successful deployment of deep convolutional neural network based SAR ATR systems in various security-sensitive applications. This paper performs a detailed adversarial robustness evaluation of deep convolutional neural network based SAR ATR models across two public available SAR target recognition datasets. For each model, seven different adversarial perturbations, ranging from gradient based optimization to self-supervised feature distortion, are generated for each testing image. Besides adversarial average recognition accuracy, feature attribution techniques have also been adopted to analyze the feature diffusion effect of adversarial attacks, which promotes the understanding of vulnerability of deep learning models.

The advancing information technology is playing more and more important role in data mining of relational database.1 The transfer and sharing of databases cause the copyright-related security threats. Database watermarking technology can effectively solve the problem with copyright protection and traceability, which has been attracting researchers' attention. In this paper, we proposed a novel, robust and reversible database watermarking technique, named histogram shifting watermarking based on random forest and genetic algorithm (RF-GAHCSW). It greatly improves the watermark capacity by means of histogram width reduction and eliminates the impact of the prediction error attack. Meanwhile, random forest algorithm is used to select important attributes for watermark embedding, and genetic algorithm is employed to find the optimal secret key for the database grouping and determine the position of watermark embedding to improve the watermark capacity and reduce data distortion. The experimental results show that the robustness of RF-GAHCSW is greatly improved, compared with the original HSW, and the distortion has little effect on the usability of database.

Due to high availability and easy accessibility of information, it has become quite difficult to assure security of data. Even though watermarking seems to be an effective solution to protect data, it is still challenging to be used with relational databases. Moreover, inserting a watermark in database may lead to distortion. As a result, the contents of database can no longer remain useful. Our proposed distortion-free watermarking approach ensures that integrity of database can be preserved by generating an image watermark from its contents. This image is registered with Certification Authority (CA) before the database is distributed for use. In case, the owner suspects any kind of tampering in the database, an image watermark is generated and compared with the registered image watermark. If both do not match, it can be concluded that the integrity of database has been compromised. Experiments are conducted on Forest Cover Type data set to localize tampering to the finest granularity. Results show that our approach can detect all types of attack with 100% accuracy.

Security and monitoring systems are more and more demanding in terms of quality, reliability and flexibility especially those dedicated to video surveillance. The quality of the acquired video signal strongly affects the performance of the high level tasks such as visual tracking, face detection and recognition. The design of a video quality assessment metric dedicated to this particular application requires a preliminary study on the common distortions encountered in video surveillance. To this end, we present in this paper a dataset dedicated to video quality assessment in the context of video surveillance. This database consists of a set of common distortions at different levels of annoyance. The subjective tests are performed using a classical pair comparison protocol with some new configurations. The subjective results obtained through the psycho-visual tests are analyzed and compared to some objective video quality assessment metrics. The preliminary results are encouraging and open a new framework for building smart video surveillance based security systems.

Video-based surveillance systems often suffer from poor-quality video in an uncontrolled environment. This may strongly affect the performance of high-level tasks such as visual tracking, abnormal event detection or more generally scene understanding and interpretation. This work aims to demonstrate the impact and the importance of video quality in video surveillance systems. Here, we focus on the most important challenges and difficulties related to the perceptual quality of the acquired or transmitted images/videos in uncontrolled environments. In this paper, we propose an architecture of a smart surveillance system that incorporates the perceptual quality of acquired scenes. We study the behaviour of some state-of-the-art video quality metrics on some original and distorted sequences from a dedicated surveillance dataset. Through this study, it has been shown that some of the state-of-the-art image/video quality metrics do not work in the context of video-surveillance. This study opens a new research direction to develop the video quality metrics in the context of video surveillance and also to propose a new quality-driven framework of video surveillance system.

With the advancement of communication technology, the interoperability of the power grid operation has improved significantly, but due to its dependence on the communication system, it is extremely vulnerable to network attacks. Among them, the false data injection attack utilizes the loophole of bad data detection in the system and attacks the state estimation system, resulting in frequent occurrence of abnormal data in the system, which brings great harm to the power grid. In view of the fact that false data injection attacks are easy to avoid traditional bad data detection methods, this paper analyzes the different situations of false data injection attacks based on the characteristics of the power grid. Firstly, it proposes to apply the distortion index method to false data injection attack detection. Experiments prove that the detection results are good and can be complementary to traditional detection methods. Then, combined with the traditional normalized residual method, this paper proposes the improved distortion index method based on the distortion index, which is good at detecting abnormal data. The use of improved distortion index method to detect false data injection attacks can make up for the defect of the lack of universality of traditional detection methods, and meet the requirements of anomaly detection efficiency. Finally, based on the MATLAB power simulation test system, experimental simulation is carried out to verify the effectiveness and universality of the proposed method for false data injection attack detection.

Binary code search has received much attention recently due to its impactful applications, e.g., plagiarism detection, malware detection and software vulnerability auditing. However, developing an effective binary code search tool is challenging due to the gigantic syntax and structural differences in binaries resulted from different compilers, compiler options and malware family. In this paper, we propose a scalable and accurate binary search engine which performs syntactic matching by combining a set of key techniques to address the challenges above. The key contribution is binary code searching technique which combined function filtering and partial trace method to match the function code relatively quick and accurate. In addition, a simhash and basic information based function filtering is proposed to dramatically reduce the irrelevant target functions. Besides, we introduce a partial trace method for matching the shortlisted function accurately. The experimental results show that our method can find similar functions, even with the presence of program structure distortion, in a scalable manner.

DeepFake using Generative Adversarial Networks (GANs) tampered videos reveals a new challenge in today's life. With the inception of GANs, generating high-quality fake videos becomes much easier and in a very realistic manner. Therefore, the development of efficient tools that can automatically detect these fake videos is of paramount importance. The proposed DeepFake detection method takes the advantage of the fact that current DeepFake generation algorithms cannot generate face images with varied resolutions, it is only able to generate new faces with a limited size and resolution, a further distortion and blur is needed to match and fit the fake face with the background and surrounding context in the source video. This transformation causes exclusive blur inconsistency between the generated face and its background in the outcome DeepFake videos, in turn, these artifacts can be effectively spotted by examining the edge pixels in the wavelet domain of the faces in each frame compared to the rest of the frame. A blur inconsistency detection scheme relied on the type of edge and the analysis of its sharpness using Haar wavelet transform as shown in this paper, by using this feature, it can determine if the face region in a video has been blurred or not and to what extent it has been blurred. Thus will lead to the detection of DeepFake videos. The effectiveness of the proposed scheme is demonstrated in the experimental results where the “UADFV” dataset has been used for the evaluation, a very successful detection rate with more than 90.5% was gained.

It is challenging to detect weak hidden information in a JPEG compressed image. In this paper, we propose a 32-layer convolutional neural networks (CNNs) with feature reuse by concatenating all features from previous layers. The proposed method can improve the flow of gradient and information, and the shared features and bottleneck layers in the proposed CNN model further reduce the number of parameters dramatically. The experimental results shown that the proposed method significantly reduce the detection error rate compared with the existing JPEG steganalysis methods, e.g. state-of-the-art XuNet method and the conventional SCA-GFR method. Compared with XuNet method and conventional method SCA-GFR in detecting J-UNIWARD at 0.1 bpnzAC (bit per non-zero AC DCT coefficient), the proposed method can reduce detection error rate by 4.33% and 6.55% respectively.

The presented method of encoding and steganographic embedding of a series of bits for the hidden message was first developed by modifying the digital platform (bases) of the elements of the image container. Unlike other methods, steganographic coding and embedding is accomplished by changing the elements of the image fragment, followed by the formation of code structures for the established structure of the digital representation of the structural elements of the image media image. The method of estimating quantitative indicators of embedded critical data is presented. The number of bits of the container for the developed method of steganographic coding and embedding of critical information is estimated. The efficiency of the presented method is evaluated and the comparative analysis of the value of the embedded digital data in relation to the method of weight coefficients of the discrete cosine transformation matrix, as well as the comparative analysis of the developed method of steganographic coding, compared with the Koch and Zhao methods to determine the embedded data resistance against attacks of various types. It is determined that for different values of the quantization coefficient, the most critical are the built-in containers of critical information, which are built by changing the part of the digital video data platform depending on the size of the digital platform and the number of bits of the built-in container.

CAPTCHA is a web-based authentication method used by websites to distinguish between humans (valid users) and bots (attackers). Audio captcha is an accessible captcha meant for the visually disabled section of users such as color-blind, blind, near-sighted users. Firstly, this paper analyzes how secure current audio captchas are from attacks using machine learning (ML) and deep learning (DL) models. Each audio captcha is made up of five, seven or ten random digits[0-9] spoken one after the other along with varying background noise throughout the length of the audio. If the ML or DL model is able to correctly identify all spoken digits and in the correct order of occurance in a single audio captcha, we consider that captcha to be broken and the attack to be successful. Throughout the paper, accuracy refers to the attack model's success at breaking audio captchas. The higher the attack accuracy, the more unsecure the audio captchas are. In our baseline experiments, we found that attack models could break audio captchas that had no background noise or medium background noise with any number of spoken digits with nearly 99% to 100% accuracy. Whereas, audio captchas with high background noise were relatively more secure with attack accuracy of 85%. Secondly, we propose that the concepts of adversarial examples algorithms can be used to create a new kind of audio captcha that is more resilient towards attacks. We found that even after retraining the models on the new adversarial audio data, the attack accuracy remained as low as 25% to 36% only. Lastly, we explore the benefits of creating adversarial audio captcha through different algorithms such as Basic Iterative Method (BIM) and deepFool. We found that as long as the attacker has less than 45% sample from each kinds of adversarial audio datasets, the defense will be successful at preventing attacks.

Robust machine learning is currently one of the most prominent topics which could potentially help shaping a future of advanced AI platforms that not only perform well in average cases but also in worst cases or adverse situations. Despite the long-term vision, however, existing studies on black-box adversarial attacks are still restricted to very specific settings of threat models (e.g., single distortion metric and restrictive assumption on target model's feedback to queries) and/or suffer from prohibitively high query complexity. To push for further advances in this field, we introduce a general framework based on an operator splitting method, the alternating direction method of multipliers (ADMM) to devise efficient, robust black-box attacks that work with various distortion metrics and feedback settings without incurring high query complexity. Due to the black-box nature of the threat model, the proposed ADMM solution framework is integrated with zeroth-order (ZO) optimization and Bayesian optimization (BO), and thus is applicable to the gradient-free regime. This results in two new black-box adversarial attack generation methods, ZO-ADMM and BO-ADMM. Our empirical evaluations on image classification datasets show that our proposed approaches have much lower function query complexities compared to state-of-the-art attack methods, but achieve very competitive attack success rates.