Algimantas Venčkauskas

Smartphone User Identity Verification Using Gait Characteristics

Smartphone-based biometrics offers a wide range of possible solutions, which could be used to authenticate users and thus to provide an extra level of security and theft prevention. We propose a method for positive identification of smartphone user’s identity using user’s gait characteristics captured by embedded smartphone sensors (gyroscopes, accelerometers). The method is based on the application of the Random Projections method for feature dimensionality reduction to just two dimensions. Then, a probability distribution function (PDF) of derived features is calculated, which is compared against known user PDF. The Jaccard distance is used to evaluate distance between two distributions, and the decision is taken based on thresholding. The results for subject recognition are at an acceptable level: we have achieved a grand mean Equal Error Rate (ERR) for subject identification of 5.7% (using the USC-HAD dataset). Our findings represent a step towards improving the performance of gait-based user identity verification technologies.

An Energy Efficient Protocol For The Internet Of Things

Abstract The Internet of Things (IoT) is a technological revolution that represents the future of computing and communications. One of the most important challenges of IoT is security: protection of data and privacy. The SSL protocol is the de-facto standard for secure Internet communications. The extra energy cost of encrypting and authenticating of the application data with SSL is around 15%. For IoT devices, where energy resources are limited, the increase in the cost of energy is a very significant factor. In this paper we present the energy efficient SSL protocol which ensures the maximum bandwidth and the required level of security with minimum energy consumption. The proper selection of the security level and CPU multiplier, can save up to 85% of the energy required for data encryption.

Generation of the Secret Encryption Key Using the Signature of the Embedded System

Program protection, programming code integrity and intellectual property protection are important problems in embedded systems. Security mechanisms for embedded systems have some specific restrictions related to limited resources, bandwidth requirements and security. In this paper we develop a secret encryption key generation algorithm by using the signature of the embedded system. We explore the qualitative characteristic of the generated keys - the entropy. Experiments showed that the generated secret keys have high entropy. DOI: http://dx.doi.org/10.5755/j01.itc.41.4.1162

Stage-Based Generative Learning Object Model to Support Automatic Content Generation and Adaptation

The paper introduces a novel Generative Learning Object (GLO) model, the Stage-Based Model (SBM) to specify the learning content. New capabilities of the model are the content automatic generation and adaptation. Externally, our model has a similar structure as the known two-level generic models (i.e. metadata and content implementation). The internal structure, however, is quite different in both parts. The use of the external parameterization technology based on pre-programming predefines the internal structure. Furthermore, the structure is derived from the initial parameterized GLO model using the refactoring tool. The technology we use in both models is based on the parameter-function relationship so that to perform manipulations on parameters. Parameters represent metadata, while the relationship implements the content variability by pre-programming the possible changes in advance so that to create the space for adaptation. The SBM implements deep internal staging by allocating parameters and functions (further objects) into predefined stages according to the given context. For example, pedagogically related parameters (objectives, teaching model, etc.) have the highest priority and appear at the top stage while the others - at the remaining stages. Typically, objects in the initial GLO specification are active, i.e. are ready to perform the prescribed role when interpreted by adequate tools. In SBM, the top stage objects are active while the remaining are passive (not able to serve the prescribed role there). The essence of the approach is the stage-based de-activation and activation of the objects within the pre-programmed specification. That ensures the automatic stage-based generation and flexibility for adaptation. In this paper, we analyze the SBM capabilities, present a case study and extended results of using the approach to the robot-oriented teaching in computer science. We also provide the pedagogical evaluation of the approach.

Model-Driven Approach for Body Area Network Application Development

This paper introduces the sensor-networked IoT model as a prototype to support the design of Body Area Network (BAN) applications for healthcare. Using the model, we analyze the synergistic effect of the functional requirements (data collection from the human body and transferring it to the top level) and non-functional requirements (trade-offs between energy-security-environmental factors, treated as Quality-of-Service (QoS)). We use feature models to represent the requirements at the earliest stage for the analysis and describe a model-driven methodology to design the possible BAN applications. Firstly, we specify the requirements as the problem domain (PD) variability model for the BAN applications. Next, we introduce the generative technology (meta-programming as the solution domain (SD)) and the mapping procedure to map the PD feature-based variability model onto the SD feature model. Finally, we create an executable meta-specification that represents the BAN functionality to describe the variability of the problem domain though transformations. The meta-specification (along with the meta-language processor) is a software generator for multiple BAN-oriented applications. We validate the methodology with experiments and a case study to generate a family of programs for the BAN sensor controllers. This enables to obtain the adequate measure of QoS efficiently through the interactive adjustment of the meta-parameter values and re-generation process for the concrete BAN application.

A comparison of authorship attribution approaches applied on the Lithuanian language

This paper reports comparative authorship attribution results obtained on the Internet comments of the morphologically complex Lithuanian language. We have explored the impact of machine learning and similarity-based approaches on the different author set sizes (containing 10, 100, and 1,000 candidate authors), feature types (lexical, morphological, and character), and feature selection techniques (feature ranking, random selection). The authorship attribution task was complicated due to the used Lithuanian language characteristics, nonnormative texts, an extreme shortness of these texts, and a large number of candidate authors. The best results were achieved with the machine learning approaches. On the larger author sets the entire feature set composed of word-level character tetra-grams demonstrated the best performance.

Covert Channel for Cluster-based File Systems Using Multiple Cover Files

Problems of sensitive information hiding in disk drives using cluster-based file systems are analyzed in this study. A new covert channel method for information hiding in disk drives is proposed and discussed. The method uses multiple cover files and is based on relative allocation of clusters of cover files in relation to one another. The experimental results presented in this paper show that the proposed method is easy to implement, provides good (for the covert channel) storage capacity and has the property of two-fold plausible deniability. The proposed covert channel method can be used for the storage of small and very sensitive information (such as passwords or encryption keys) on removable disk drives. DOI: http://dx.doi.org/10.5755/j01.itc.42.3.3328

Evaluation of Open Source Server-Side XSS Protection Solutions

Web protection against XSS attacks is an indispensable tool for implementing reliable online systems. XSS attacks can be used for various malicious actions and stealing important information. Protection may be implemented both on user computer and on server side. In this work we have analyzed the server side protection solutions. These solutions must ensure appropriate level of security and at the same time should not considerably increase page response time. The aim of this paper is to determine the most effective and safe free tools for protection against XSS attacks for web pages created using PHP, ASP.NET and Java technologies.

A Lightweight Protocol for Secure Video Streaming

The Internet of Things (IoT) introduces many new challenges which cannot be solved using traditional cloud and host computing models. A new architecture known as fog computing is emerging to address these technological and security gaps. Traditional security paradigms focused on providing perimeter-based protections and client/server point to point protocols (e.g., Transport Layer Security (TLS)) are no longer the best choices for addressing new security challenges in fog computing end devices, where energy and computational resources are limited. In this paper, we present a lightweight secure streaming protocol for the fog computing “Fog Node-End Device” layer. This protocol is lightweight, connectionless, supports broadcast and multicast operations, and is able to provide data source authentication, data integrity, and confidentiality. The protocol is based on simple and energy efficient cryptographic methods, such as Hash Message Authentication Codes (HMAC) and symmetrical ciphers, and uses modified User Datagram Protocol (UDP) packets to embed authentication data into streaming data. Data redundancy could be added to improve reliability in lossy networks. The experimental results summarized in this paper confirm that the proposed method efficiently uses energy and computational resources and at the same time provides security properties on par with the Datagram TLS (DTLS) standard.

Open class authorship attribution of lithuanian internet comments using one-class classifier

Internet can be misused by cyber criminals as a platform to conduct illegitimate activities (such as harassment, cyber bullying, and incitement of hate or violence) anonymously. As a result, authorship analysis of anonymous texts in Internet (such as emails, forum comments) has attracted significant attention in the digital forensic and text mining communities. The main problem is a large number of possible of authors, which hinders the effective identification of a true author. We interpret open class author attribution as a process of expert recommendation where the decision support system returns a list of suspected authors for further analysis by forensics experts rather than a single prediction result, thus reducing the scale of the problem. We describe the task formally and present algorithms for constructing the suspected author list. For evaluation we propose using a simple Winner-Takes-All (WTA) metric as well as a set of gain-discount model based metrics from the information retrieval domain (mean reciprocal rank, discounted cumulative gain and rank-biased precision). We also propose the List Precision (LP) metric as an extension of WTA for evaluating the usability of the suspected author list. For experiments, we use our own dataset of Internet comments in Lithuanian language and consider the use of language-specific (Lithuanian) lexical features together with general lexical features derived from English language. For classification we use one-class Support Vector Machine (SVM) classifier. The results of experiments show that the usability of open class author attribution can be improved considerably by using a set of language-specific lexical features together with general lexical features, while the proposed method can be used to reduce the number of suspected authors thus alleviating the work of forensic linguists.