Biblio

Twitter text analysis is quite useful in analysing emotions, sentiments and feedbacks of consumers on products and services. This helps the service providers and the manufacturers to improve their products and services, address serious issues before they lead to a crisis and improve business acumen. Twitter texts also form a data source for various research studies. They are used in topic analysis, sentiment analysis, content analysis and thematic analysis. In this paper, we present a pipeline for searching, analysing and visualizing the text analytics features of twitter texts using web APIs. It allows to build a simple yet powerful twitter text analytics tool for researchers and other interested users.

Nowadays, video surveillance systems are part of our daily life, because of their role in ensuring the security of goods and people this generates a huge amount of video data. Thus, several research works based on the ontology paradigm have tried to develop an efficient system to index and search precisely a very large volume of videos. Due to their semantic expressiveness, ontologies are undoubtedly very much in demand in recent years in the field of video surveillance to overcome the problem of the semantic gap between the interpretation of the data extracted from the low level and the high-level semantics of the video. Despite its good expressiveness of semantics, a classical ontology may not be sufficient for good handling of uncertainty, which is however commonly present in the video surveillance domain, hence the need to consider a new ontological approach that will better represent uncertainty. Fuzzy logic is recognized as a powerful tool for dealing with vague, incomplete, imperfect, or uncertain data or information. In this work, we develop a new ontological approach based on fuzzy logic. All the relevant fuzzy concepts such as Video\_Objects, Video\_Events, Video\_Sequences, that could appear in a video surveillance domain are well represented with their fuzzy Ontology DataProperty and the fuzzy relations between them (Ontology ObjectProperty). To achieve this goal, the new fuzzy video surveillance ontology is implemented using the fuzzy ontology web language 2 (fuzzy owl2) which is an extension of the standard semantic web language, ontology web language 2 (owl2).

In new technological world pervasive computing plays the important role in data computing and communication. The pervasive computing provides the mobile environment for decentralized computational services at anywhere, anytime at any context and location. Pervasive computing is flexible and makes portable devices and computing surrounded us as part of our daily life. Devices like Laptop, Smartphones, PDAs, and any other portable devices can constitute the pervasive environment. These devices in pervasive environments are worldwide and can receive various communications including audio visual services. The users and the system in this pervasive environment face the challenges of user trust, data privacy and user and device node identity. To give the feasible determination for these challenges. This paper aims to propose a dynamic learning in pervasive computing environment refer the challenges proposed efficient security model (ESM) for trustworthy and untrustworthy attackers. ESM model also compared with existing generic models; it also provides better accuracy rate than existing models.

Implementing the color images encryption is a challenging field of the research for IOT applications. An exponential growth in imaging cameras in IOT uses makes it critical to design the robust image security algorithms. It is also observed that performance of existing encryption methods degrades under the presence of noisy environments. This is the major concern of evaluating the encryption method in this paper. The prime concern of this paper is to design the fast efficient color images encryption algorithm by designing an efficient and robustness RSA encryption algorithm. Method takes the advantage of both preprocessing and the Gaussian pyramid (GP) approach for encryption. To improve the performance it is proposed to use the LAB color space and implement the RSA encryption on luminance (L) component using the GP domain. The median filter and image sharpening is used for preprocessing. The goal is to improve the performance under highly noisy imaging environment. The performance is compared based on the crypto weights and on the basis of visual artifacts and entropy analysis. The decrypted outputs are again converted to color image output. Using the LAB color space is expected to improve the entropy performance of the image. Result of proposed encryption method is evaluated under the different types of the noisy attacks over the color images and also performance is compared with state of art encryption methods. Significant improvement speed of the algorithm is compared in terms of the elapsed time

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.

Nowadays, developing an intelligent system for segmenting the moving object from the background is essential task for video surveillance applications. Recently, a deep learning segmentation algorithm composed of encoder CNN, a Feature Pooling Module and a decoder CNN called FgSegNET\_S has been proposed. It is capable to train the model using few training examples. FgSegNET\_S is relying only on the spatial information while it is fundamental to include temporal information to distinguish if an object is moving or not. In this paper, an improved version known as (T\_FgSegNET\_S) is proposed by using the subtracted images from the initial background as input. The proposed approach is trained and evaluated using two publicly available infrared datasets: remote scene infrared videos captured by medium-wave infrared (MWIR) sensors and the Grayscale Thermal Foreground Detection (GTFD) dataset. The performance of network is evaluated using precision, recall, and F-measure metrics. The experiments show improved results, especially when compared to other state-of-the-art methods.

The main objective of Human Activity Recognition (HAR) is to detect various activities in video frames. Video surveillance is an import application for various security reasons, therefore it is essential to classify activities as usual and unusual. This paper implements the deep learning model that has the ability to classify and localize the activities detected using a Single Shot Detector (SSD) algorithm with a bounding box, which is explicitly trained to detect usual and unusual activities for security surveillance applications. Further this model can be deployed in public places to improve safety and security of individuals. The SSD model is designed and trained using transfer learning approach. Performance evaluation metrics are visualised using Tensor Board tool. This paper further discusses the challenges in real-time implementation.

This Wiping Tool is an anti-forensic tool that is built to wipe data permanently from laptop's storage. This tool is capable to ensure the data from being recovered with any recovery tools. The objective of building this wiping tool is to maintain the confidentiality and integrity of the data from unauthorized access. People tend to delete the file in normal way, however, the file face the risk of being recovered. Hence, the integrity and confidentiality of the deleted file cannot be protected. Through wiping tools, the files are overwritten with random strings to make the files no longer readable. Thus, the integrity and the confidentiality of the file can be protected. Regarding wiping tools, nowadays, lots of wiping tools face issue such as data breach because the wiping tools are unable to delete the data permanently from the devices. This situation might affect their main function and a threat to their users. Hence, a new wiping tool is developed to overcome the problem. A new wiping tool named Data Wiping tool is applying two wiping techniques. The first technique is Randomized Data while the next one is enhancing wiping technique, known as ByteEditor. ByteEditor is a combination of two different techniques, byte editing and byte deletion. With the implementation of Object-Oriented methodology, this wiping tool is built. This methodology consists of analyzing, designing, implementation and testing. The tool is analyzed and compared with other wiping tools before the designing of the tool start. Once the designing is done, implementation phase take place. The code of the tool is created using Visual Studio 2010 with C\# language and being tested their functionality to ensure the developed tool meet the objectives of the project. This tool is believed able to contribute to the development of wiping tools and able to solve problems related to other wiping tools.

With the maturity of Internet of things technology, massive data transmission has become the focus of research. In order to solve the problem of low speed of traditional hybrid data fast distribution algorithm for wireless sensor networks, a hybrid data fast distribution algorithm for wireless sensor networks based on visual Internet of things is designed. The logic structure of mixed data input gate in wireless sensor network is designed through the visual Internet of things. The objective function of fast distribution of mixed data in wireless sensor network is proposed. The number of copies of data to be distributed is dynamically calculated and the message deletion strategy is determined. Then the distribution parameters are calibrated, and the fitness ranking is performed according to the distribution quantity to complete the algorithm design. The experimental results show that the distribution rate of the designed algorithm is significantly higher than that of the control group, which can solve the problem of low speed of traditional data fast distribution algorithm.

The continuous increase in sophistication of threat actors over the years has made the use of actionable threat intelligence a critical part of the defence against them. Such Cyber Threat Intelligence is published daily on several online sources, including vulnerability databases, CERT feeds, and social media, as well as on forums and web pages from the Surface and the Dark Web. Named Entity Recognition (NER) techniques can be used to extract the aforementioned information in an actionable form from such sources. In this paper we investigate how the latest advances in the NER domain, and in particular transformer-based models, can facilitate this process. To this end, the dataset for NER in Threat Intelligence (DNRTI) containing more than 300 pieces of threat intelligence reports from open source threat intelligence websites is used. Our experimental results demonstrate that transformer-based techniques are very effective in extracting cybersecurity-related named entities, by considerably outperforming the previous state- of-the-art approaches tested with DNRTI.

The dark web has become a major trading platform for cybercriminals, with its anonymity and encrypted content nature make it possible to exchange hacked information and sell illegal goods without being traced. The types of items traded on the dark web have increased with the number of users and demands. In recent years, in addition to the main items sold in the past, including drugs, firearms and child pornography, a growing number of cybercriminals are targeting various types of private information, including different types of account data, identity information and visual data etc. This paper will further discuss the issue of threat detection in the dark web by reviewing the past literature on the subject. An approach is also proposed to identify criminals who commit crimes offline or on the surface network by using private information purchased from the dark web and the original sources of information on the dark web by building a database based on historical victim records for keyword matching and traffic analysis.

Nowadays, peoples are very concerned about their data privacy. Hence, all the current security methods should be improved to stay relevant in this fast-growing technology world. Visual Cryptography (VC) is a cryptographic technique that using the image processing method. The implementation of VC can be varying and flexible to be applied to the system that requires an extra security precaution as it is one of the effective solutions in securing the data exchange between two or more parties. The main purpose of the development of V-CRYPT System is to improve the current VC technique and make it more complex in the encryption and decryption process. V-CRYPT system will let the user enter the key, then select the image that they want to encrypt, and the system will split the image into four shares: share0, share1, share2, share3. Each pixel of the image will be splatted into a smaller block of subpixels in each of the four shares and encrypted as two subpixels in each of the shares. The decryption will work only when the user selects all the shares, and the correct text key is entered. The system will superimpose all the shares and producing one perfect image. If the incorrect key is entered, the resulted image will be unidentified. The results show that V- CRYPT is a valuable alternative to existing methods where its security level is higher in terms of adding a secure key and complexity.

This paper proposes a digital image compression-encryption scheme based on the theory of singular value decomposition, multiple chaos and Beta function, which uses SVD to compress the digital image and utilizes three way protections for encryption viz. logistic and Arnold map along with the beta function. The algorithm has three advantages: First, the compression scheme gives the freedom to a user so that one can select the desired compression level according to the application with the help of singular value. Second, it includes a confusion mechanism wherein the pixel positions of image are scrambled employing Cat Map. The pixel location is shuffled, resulting in a cipher text image that is safe for communication. Third the key is generated with the help of logistic map which is nonlinear and chaotic in nature therefore highly secured. Fourth the beta function used for encryption is symmetric in nature which means the order of its parameters does not change the outcome of the operation, meaning faithful reconstruction of an image. Thus, the algorithm is highly secured and also saving the storage space as well. The experimental results show that the algorithm has the advantages of faithful reconstruction with reasonable PSNR on different singular values.

This paper builds an Augmented Reality Sandbox (AR Sandbox) system based on augmented reality technology, and performs a 3D reconstruction for the sandbox terrain using the depth sensor Microsoft Kinect in the AR Sandbox, as an entry point to pave the way for later development of related metaverse applications, such as the metaverse architecting and visual interactive modeling. The innovation of this paper is that for the AR Sandbox scene, a 3D reconstruction method based on depth sensor is proposed, which can automatically cut off the edge of the sandbox table in Kinect field of view, and accurately and completely reconstruct the sandbox terrain in Matlab.

Style transfer means using a neural network to extract the content of one image and the style of the other image. The two are combined to get the final result, broadly applied in social communication, animation production, entertainment items. Using style transfer, users can share and exchange images; painters can create specific art styles more readily with less creation cost and production time. Therefore, style transfer is widely concerned recently due to its various and valuable applications. In the past few years, the paper reviews style transfer and chooses three representative works to analyze in detail and contrast with each other, including StyleGAN, CycleGAN, and TL-GAN. Moreover, what function an ideal model of style transfer should realize is discussed. Compared with such a model, potential problems and prospects of different methods to achieve style transfer are listed. A couple of solutions to these drawbacks are given in the end.

The use of static code analysis tools for security audits can be time consuming, as the many existing tools focus on different aspects and therefore development teams often use several of these tools to keep code quality high and prevent security issues. Displaying the results of multiple tools, such as code smells and security warnings, in a unified interface can help developers get a better overview and prioritize upcoming work. We present visualizations and a dashboard that interactively display results from static code analysis for “interesting” commits during development. With this, we aim to provide an effective visual analytics tool for code security analysis results.

Clustering artworks is difficult for several reasons. On the one hand, recognizing meaningful patterns in accordance with domain knowledge and visual perception is extremely difficult. On the other hand, applying traditional clustering and feature reduction techniques to the highly dimensional pixel space can be ineffective. To address these issues, we propose to use a deep convolutional embedding model for digitized painting clustering, in which the task of mapping the raw input data to an abstract, latent space is jointly optimized with the task of finding a set of cluster centroids in this latent feature space. Quantitative and qualitative experimental results show the effectiveness of the proposed method. The model is also capable of outperforming other state-of-the-art deep clustering approaches to the same problem. The proposed method can be useful for several art-related tasks, in particular visual link retrieval and historical knowledge discovery in painting datasets.

The network output activation values for a given input can be employed to produce a sorted ranking. Adversarial attacks typically generate the least amount of perturbation required to change the classifier label. In that sense, generated adversarial attack perturbation only affects the output in the 1st sorted ranking position. We argue that meaningful information about the adversarial examples i.e., their original labels, is still encoded in the network output ranking and could potentially be extracted, using rule-based reasoning. To this end, we introduce a novel adversarial attack methodology inspired by the K-anonymity principles, that generates adversarial examples that are not only misclassified, but their output sorted ranking spreads uniformly along K different positions. Any additional perturbation arising from the strength of the proposed objectives, is regularized by a visual similarity-based term. Experimental results denote that the proposed approach achieves the optimization goals inspired by K-anonymity with reduced perturbation as well.

Due to respectively limited training data, different entities addressing the same vision task based on certain sensitive images may not train a robust deep network. This paper introduces a new vision task where various entities share task-specific image data to enlarge each other's training data volume without visually disclosing sensitive contents (e.g. illegal images). Then, we present a new structure-based training regime to enable different entities learn task-specific and reconstruction-proof image representations for image data sharing. Specifically, each entity learns a private Deep Poisoning Module (DPM) and insert it to a pre-trained deep network, which is designed to perform the specific vision task. The DPM deliberately poisons convolutional image features to prevent image reconstructions, while ensuring that the altered image data is functionally equivalent to the non-poisoned data for the specific vision task. Given this equivalence, the poisoned features shared from one entity could be used by another entity for further model refinement. Experimental results on image classification prove the efficacy of the proposed method.

In these days, malware attacks target different kinds of devices as IoT, mobiles, servers even the cloud. It causes several hardware damages and financial losses especially for big companies. Malware attacks represent a serious issue to cybersecurity specialists. In this paper, we propose a new approach to detect unknown malware families based on machine learning classification and visualization technique. A malware binary is converted to grayscale image, then for each image a GIST descriptor is used as input to the machine learning model. For the malware classification part we use 3 machine learning algorithms. These classifiers are so efficient where the highest precision reach 98%. Once we train, test and evaluate models we move to simulate 2 new malware families. We do not expect a good prediction since the model did not know the family; however our goal is to analyze the behavior of our classifiers in the case of new family. Finally, we propose an approach using a filter to know either the classification is normal or it's a zero-day malware.

The proliferation of malware attacks causes a hindrance to cybersecurity thus, posing a significant threat to our devices. The variety and number of both known as well as unknown malware makes it difficult to detect it. Research suggests that the ramifications of malware are only becoming worse with time and hence malware analysis becomes crucial. This paper proposes a visual malware classification technique to convert malware executables into their visual representations and obtain grayscale images of malicious files. These grayscale images are then used to classify malicious files into their respective malware families by passing them through deep convolutional neural networks (CNN). As part of deep CNN, we use various ImageNet models and compare their performance.

A malware detection model based on semi-supervised learning is proposed in the paper. Our model includes mainly three parts: malware visualization, feature extraction, and classification. Firstly, the malware visualization converts malware into grayscale images; then the features of the images are extracted to reflect the coding patterns of malware; finally, a collaborative learning model is applied to malware detections using both labeled and unlabeled software samples. The proposed model was evaluated based on two commonly used benchmark datasets. The results demonstrated that compared with traditional methods, our model not only reduced the cost of sample labeling but also improved the detection accuracy through incorporating unlabeled samples into the collaborative learning process, thereby achieved higher classification performance.

Advances in artificial intelligence, machine learning, and robotics have profoundly impacted the field of autonomous navigation and driving. However, sensor spoofing attacks can compromise critical components and the control mechanisms of mobile robots. Therefore, understanding vulnerabilities in autonomous driving and developing countermeasures remains imperative for the safety of unmanned vehicles. Hence, we demonstrate cross-validation techniques for detecting spoofing attacks on the sensor data in autonomous driving in this work. First, we discuss how visual and inertial odometry (VIO) algorithms can provide a root-of-trust during navigation. Then, we develop examples for sensor data spoofing attacks using the open-source driving dataset. Next, we design an attack detection technique using VIO algorithms that cross-validates the navigation parameters using the IMU and the visual data. Following, we consider hardware-dependent attack survival mechanisms that support an autonomous system during an attack. Finally, we also provide an example of spoofing survival technique using on-board hardware oscillators. Our work demonstrates the applicability of classical mobile robotics algorithms and hardware security primitives in defending autonomous vehicles from targeted cyber attacks.

Security is a major thing to focus on during this modern era as it is very important to secure your surroundings for the well being of oneself and his family, But there are many drawbacks of using conventional security surveillance cameras as they have to be set in a particular angle for good visual and they do not cover a large area, conventional security cameras can only be used from a particular device and cannot alert the user during an unforeseen circumstance. Hence we require a much more efficient device for better security a web controlled surveillance robot is much more practical device to be used compared to conventional security surveillance, this system needs a single camera to perform its operation and the user can monitor a wide range of area, any device with a wireless connection to the internet can be used to operate this device. This robot can move to any location within the range of the network and can be accessed globally from anywhere and as it uses only one camera to secure a large area it is also cost-efficient. At the core of the system lies Raspberry-pi which is responsible for all the operation of the system and the size of the device can be engineered according to the area it is to be used.

Deep learning (DL) classifiers are often unstable in that they may change significantly when retested on perturbed images or low quality images. This paper adds to the fundamental body of work on the robustness of DL classifiers. We introduce a new two-dimensional benchmarking matrix to evaluate robustness of DL classifiers, and we also innovate a four-quadrant statistical visualization tool, including minimum accuracy, maximum accuracy, mean accuracy, and coefficient of variation, for benchmarking robustness of DL classifiers. To measure robust DL classifiers, we create comprehensive 69 benchmarking image sets, including a clean set, sets with single factor perturbations, and sets with two-factor perturbation conditions. After collecting experimental results, we first report that using two-factor perturbed images improves both robustness and accuracy of DL classifiers. The two-factor perturbation includes (1) two digital perturbations (salt & pepper noise and Gaussian noise) applied in both sequences, and (2) one digital perturbation (salt & pepper noise) and a geometric perturbation (rotation) applied in both sequences. All source codes, related image sets, and results are shared on the GitHub website at https://github.com/caperock/robustai to support future academic research and industry projects.