Description:
Industrial Internet of Things (IIoT) deploys edge devices to act as intermediaries between sensors and actuators and application servers or cloud services. Machine learning models have been widely used to thwart malware attacks in such edge devices. However, these models are vulnerable to adversarial attacks where attackers craft adversarial samples by introducing small perturbations to malware samples to fool a classifier to misclassify them as benign applications. Literature on deep learning networks proposes adversarial retraining as a defense mechanism where adversarial samples are combined with legitimate samples to retrain the classifier. However, existing works select such adversarial samples in a random fashion which degrades the classifier's performance. This work proposes two novel approaches for selecting adversarial samples to retrain a classifier. One, based on the distance from malware cluster center, and the other, based on a probability measure derived from a kernel based learning (KBL). Our experiments show that both of our sample selection methods outperform the random selection method and the KBL selection method improves detection accuracy by 6%. Also, while existing works focus on deep neural networks with respect to adversarial retraining, we additionally assess the impact of such adversarial samples on other classifiers and our proposed selective adversarial retraining approaches show similar performance improvement for these classifiers as well. The outcomes from the study can assist in designing robust security systems for IIoT applications.
Description:
Industrial Internet of Things (IIoT) deploys edge devices to act as intermediaries between sensors and actuators and application servers or cloud services. Machine learning models have been widely used to thwart malware attacks in such edge devices. However, these models are vulnerable to adversarial attacks where attackers craft adversarial samples by introducing small perturbations to malware samples to fool a classifier to misclassify them as benign applications. Literature on deep learning networks proposes adversarial retraining as a defense mechanism where adversarial samples are combined with legitimate samples to retrain the classifier. However, existing works select such adversarial samples in a random fashion which degrades the classifier's performance. This work proposes two novel approaches for selecting adversarial samples to retrain a classifier. One, based on the distance from malware cluster center, and the other, based on a probability measure derived from a kernel based learning (KBL). Our experiments show that both of our sample selection methods outperform the random selection method and the KBL selection method improves detection accuracy by 6%. Also, while existing works focus on deep neural networks with respect to adversarial retraining, we additionally assess the impact of such adversarial samples on other classifiers and our proposed selective adversarial retraining approaches show similar performance improvement for these classifiers as well. The outcomes from the study can assist in designing robust security systems for IIoT applications.
Description:
Internet of Things (IoT) image sensors, social media, and smartphones generate huge volumes of digital images every day. Easy availability and usability of photo editing tools have made forgery attacks, primarily splicing and copy-move attacks, effortless, causing cybercrimes to be on the rise. While several models have been proposed in the literature for detecting these attacks, the robustness of those models has not been investigated when (i) a low number of tampered images are available for model building or (ii) images from IoT sensors are distorted due to image rotation or scaling caused by unwanted or unexpected changes in sensors' physical set-up. Moreover, further improvement in detection accuracy is needed for real-word security management systems. To address these limitations, in this paper, an innovative image forgery detection method has been proposed based on Discrete Cosine Transformation (DCT) and Local Binary Pattern (LBP) and a new feature extraction method using the mean operator. First, images are divided into non-overlapping fixed size blocks and 2D block DCT is applied to capture changes due to image forgery. Then LBP is applied to the magnitude of the DCT array to enhance forgery artifacts. Finally, the mean value of a particular cell across all LBP blocks is computed, which yields a fixed number of features and presents a more computationally efficient method. Using Support Vector Machine (SVM), the proposed method has been extensively tested on four well known publicly available gray scale and color image forgery datasets, and additionally on an IoT based image forgery dataset that we built. Experimental results reveal the superiority of our proposed method over recent state-of-the-art methods in terms of widely used performance metrics and computational time and demonstrate robustness against low availability of forged training samples.
Description:
This research was funded by Research Priority Area (RPA) scholarship of Federation University Australia.
Description:
Internet of Things (IoT) image sensors, social media, and smartphones generate huge volumes of digital images every day. Easy availability and usability of photo editing tools have made forgery attacks, primarily splicing and copy-move attacks, effortless, causing cybercrimes to be on the rise. While several models have been proposed in the literature for detecting these attacks, the robustness of those models has not been investigated when (i) a low number of tampered images are available for model building or (ii) images from IoT sensors are distorted due to image rotation or scaling caused by unwanted or unexpected changes in sensors' physical set-up. Moreover, further improvement in detection accuracy is needed for real-word security management systems. To address these limitations, in this paper, an innovative image forgery detection method has been proposed based on Discrete Cosine Transformation (DCT) and Local Binary Pattern (LBP) and a new feature extraction method using the mean operator. First, images are divided into non-overlapping fixed size blocks and 2D block DCT is applied to capture changes due to image forgery. Then LBP is applied to the magnitude of the DCT array to enhance forgery artifacts. Finally, the mean value of a particular cell across all LBP blocks is computed, which yields a fixed number of features and presents a more computationally efficient method. Using Support Vector Machine (SVM), the proposed method has been extensively tested on four well known publicly available gray scale and color image forgery datasets, and additionally on an IoT based image forgery dataset that we built. Experimental results reveal the superiority of our proposed method over recent state-of-the-art methods in terms of widely used performance metrics and computational time and demonstrate robustness against low availability of forged training samples.
Description:
This research was funded by Research Priority Area (RPA) scholarship of Federation University Australia.
Description:
Confidentiality of digital images is an important requirement for many multimedia applications and services. To maintain confidentiality, encryption of digital images is essential. Digital images are usually very large and encrypting such bulky data induces many performance overheads, which can be too expensive for real-time applications in resource constrained environments. In this paper, we propose a chaotic image encryption scheme which satisfies the need for both light-weightedness and security. To justify the security and efficiency, the new cipher was evaluated using a series of statistical tests. These tests included a visual testing and a histogram analysis, a randomness analysis, a correlation analysis, an entropy analysis and an image encryption quality analysis. Based on all analyses and experimental results, it is concluded that the proposed scheme is effective, efficient and trustworthy and therefore can be adopted for image encryption.
Description:
Confidentiality of digital images is an important requirement for many multimedia applications and services. To maintain confidentiality, encryption of digital images is essential. Digital images are usually very large and encrypting such bulky data induces many performance overheads, which can be too expensive for real-time applications in resource constrained environments. In this paper, we propose a chaotic image encryption scheme which satisfies the need for both light-weightedness and security. To justify the security and efficiency, the new cipher was evaluated using a series of statistical tests. These tests included a visual testing and a histogram analysis, a randomness analysis, a correlation analysis, an entropy analysis and an image encryption quality analysis. Based on all analyses and experimental results, it is concluded that the proposed scheme is effective, efficient and trustworthy and therefore can be adopted for image encryption.
Description:
The development of a complex and dynamic system such as the energy sector requires a comprehensive understanding of its constituent components and their interactions, and thus requires approaches that can adapt to the dynamic complexity in systems. Previous efforts mainly used reductionist approaches, which examine the components of the system in isolation, neglecting their interdependent nature. Such approaches reduce our ability to understand the system and/or mitigate undesirable outcomes. We adopt a system dynamics approach to construct an integrated model for analysing the behaviour of the energy sector. Although the Australian energy sector is used as a case study, the model can be applied in other context elsewhere around the world The results indicate that the current trajectory of the Australian energy sector is unsustainable and growth is not being controlled. Limits to growth are fast approaching due to excessive fossil fuel extraction, high emissions and high energy dependency. With the current growth, Australia's global CO2 emissions footprint will increase to unprecedented levels reaching 12% by 2030 (9.5% for exports and 2.5% for domestic). Oil dependency will account for 43% and 47% of total consumption by 2030 and 2050. By 2032, coal will be the only fossil fuel resource available in Australia. Expansion of investment in coal and gas production is a large risk.
Description:
The development of a complex and dynamic system such as the energy sector requires a comprehensive understanding of its constituent components and their interactions, and thus requires approaches that can adapt to the dynamic complexity in systems. Previous efforts mainly used reductionist approaches, which examine the components of the system in isolation, neglecting their interdependent nature. Such approaches reduce our ability to understand the system and/or mitigate undesirable outcomes. We adopt a system dynamics approach to construct an integrated model for analysing the behaviour of the energy sector. Although the Australian energy sector is used as a case study, the model can be applied in other context elsewhere around the world The results indicate that the current trajectory of the Australian energy sector is unsustainable and growth is not being controlled. Limits to growth are fast approaching due to excessive fossil fuel extraction, high emissions and high energy dependency. With the current growth, Australia's global CO2 emissions footprint will increase to unprecedented levels reaching 12% by 2030 (9.5% for exports and 2.5% for domestic). Oil dependency will account for 43% and 47% of total consumption by 2030 and 2050. By 2032, coal will be the only fossil fuel resource available in Australia. Expansion of investment in coal and gas production is a large risk.
Description:
Web applications have become increasingly popular due to their potential for businesses' high revenue gain through global reach. Along with these opportunities, also come challenges in terms of Web application security. The increased rise in the number of datadriven applications has also seen an increased rise in their systematic attacks. Cyberattacks exploit Web application vulnerabilities. Attack trends show a major increase in Web application vulnerabilities caused by improper implementation of information-flow control methods and they account for more than 50% of all Web application vulnerabilities found in the year 2013. Static code analysis using methods of information-flow control is a widely acknowledged technique to secure Web applications. Whilst this technique has been found to be both very effective and efficient in finding Web application vulnerabilities, specific tools are highly dependent on the programming language. This thesis leverages Semantic Web technologies in order to offer a common language through source code represented using the Resource Description Framework format, whereby reasoning can be applied to securely test Web applications. In this thesis, we present a framework that extracts source code facts from various programming languages at a variable-level of granularity using Abstract Syntax Trees (ASTs) generated using language grammars and the ANTLR parser generator. The methodology for detecting Web application vulnerabilities implements three phases: entry points identification, tracing information-flow and vulnerability detection using the Jena framework inference mechanism and rules describing patterns of source code. The approach discussed in this thesis is found to be effective and practical in finding Web application vulnerabilities with the limitation that it can only detect patterns that are used as training data or very similar patterns. False positives are caused by limitations of the language grammar, but they do not affect the accuracy of the security vulnerability detection method in identifying the correct Web application vulnerability.
Description:
Web applications have become increasingly popular due to their potential for businesses' high revenue gain through global reach. Along with these opportunities, also come challenges in terms of Web application security. The increased rise in the number of datadriven applications has also seen an increased rise in their systematic attacks. Cyberattacks exploit Web application vulnerabilities. Attack trends show a major increase in Web application vulnerabilities caused by improper implementation of information-flow control methods and they account for more than 50% of all Web application vulnerabilities found in the year 2013. Static code analysis using methods of information-flow control is a widely acknowledged technique to secure Web applications. Whilst this technique has been found to be both very effective and efficient in finding Web application vulnerabilities, specific tools are highly dependent on the programming language. This thesis leverages Semantic Web technologies in order to offer a common language through source code represented using the Resource Description Framework format, whereby reasoning can be applied to securely test Web applications. In this thesis, we present a framework that extracts source code facts from various programming languages at a variable-level of granularity using Abstract Syntax Trees (ASTs) generated using language grammars and the ANTLR parser generator. The methodology for detecting Web application vulnerabilities implements three phases: entry points identification, tracing information-flow and vulnerability detection using the Jena framework inference mechanism and rules describing patterns of source code. The approach discussed in this thesis is found to be effective and practical in finding Web application vulnerabilities with the limitation that it can only detect patterns that are used as training data or very similar patterns. False positives are caused by limitations of the language grammar, but they do not affect the accuracy of the security vulnerability detection method in identifying the correct Web application vulnerability.
Description:
Security is an integral component of the relationships that shape the lives of students hosted by foreign countries. International student security became a source of contention between nations in 2008 when China's Government charged that Australia was failing to adequately provide for the safety of Chinese students. We discuss Beck's theorisation of the ‘risk society’ and Clements' notion of balancing risk with opportunity. We draw on interviews undertaken in Beijing to highlight the importance of student security and the need for hosts to accord attention to the views of parents as well as students. Our findings reveal that parents play a major role in the study location decision, tend to accord security greater weight than do students when location is being debated within the family, utilise a range of strategies to keep students secure and believe host governments and institutions have primary responsibility for student well being.
Description:
Due to the widespread application of augmented and virtual environments, the research into 3D content protection is fundamentally important. To maintain confidentiality, encryption of 3D content, including the 3D objects and texture images, is essential. In this paper, a novel texture encryption scheme is proposed which complements the existing 3D object encryption methods. The proposed method encrypts texture images by bit masking and a permutation procedure using the Salsa20/12 stream cipher. The method is lightweight and satisfies the security requirement. It also prevents the partial disclosure of the encrypted 3D surface geometry by protecting the texture patterns from being partially leaked. The scheme has a better speed-security profile than the full encryption and the selective (4most significant bit-plane) encryption by 128-bit AES. The encryption schemes are implemented and tested with 500 sample texture images. The experimental results show that the scheme has a better encryption performance compared to the full/selective encryption by 128-bit AES.
Description:
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Description:
Due to the widespread application of augmented and virtual environments, the research into 3D content protection is fundamentally important. To maintain confidentiality, encryption of 3D content, including the 3D objects and texture images, is essential. In this paper, a novel texture encryption scheme is proposed which complements the existing 3D object encryption methods. The proposed method encrypts texture images by bit masking and a permutation procedure using the Salsa20/12 stream cipher. The method is lightweight and satisfies the security requirement. It also prevents the partial disclosure of the encrypted 3D surface geometry by protecting the texture patterns from being partially leaked. The scheme has a better speed-security profile than the full encryption and the selective (4most significant bit-plane) encryption by 128-bit AES. The encryption schemes are implemented and tested with 500 sample texture images. The experimental results show that the scheme has a better encryption performance compared to the full/selective encryption by 128-bit AES.
Description:
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Description:
The revolution in online criminal activities and malicious software (malware) has posed a serious challenge in malware forensics. Malicious attacks have become more organized and purposefully directed. With cybercrimes escalating to great heights in quantity as well as in sophistication and stealth, the main challenge is to detect hidden and obfuscated malware. Malware authors use a variety of obfuscation methods and specialized stealth techniques of information hiding to embed malicious code, to infect systems and to thwart any attempt to detect them, specifically with the use of commercially available anti-malware engines. This has led to the situation of zero-day attacks, where malware inflict systems even with existing security measures. The aim of this thesis is to address this situation by proposing a variety of novel digital forensic and data mining techniques to automatically detect hidden and obfuscated malware. Anti-malware engines use signature matching to detect malware where signatures are generated by human experts by disassembling the file and selecting pieces of unique code. Such signature based detection works effectively with known malware but performs poorly with hidden or unknown malware. Code obfuscation techniques, such as packers, polymorphism and metamorphism, are able to fool current detection techniques by modifying the parent code to produce offspring copies resulting in malware that has the same functionality, but with a different structure. These evasion techniques exploit the drawbacks of traditional malware detection methods, which take current malware structure and create a signature for detecting this malware in the future. However, obfuscation techniques aim to reduce vulnerability to any kind of static analysis to the determent of any reverse engineering process. Furthermore, malware can be hidden in file system slack space, inherent in NTFS file system based partitions, resulting in malware detection that even more difficult.
Description:
The revolution in online criminal activities and malicious software (malware) has posed a serious challenge in malware forensics. Malicious attacks have become more organized and purposefully directed. With cybercrimes escalating to great heights in quantity as well as in sophistication and stealth, the main challenge is to detect hidden and obfuscated malware. Malware authors use a variety of obfuscation methods and specialized stealth techniques of information hiding to embed malicious code, to infect systems and to thwart any attempt to detect them, specifically with the use of commercially available anti-malware engines. This has led to the situation of zero-day attacks, where malware inflict systems even with existing security measures. The aim of this thesis is to address this situation by proposing a variety of novel digital forensic and data mining techniques to automatically detect hidden and obfuscated malware. Anti-malware engines use signature matching to detect malware where signatures are generated by human experts by disassembling the file and selecting pieces of unique code. Such signature based detection works effectively with known malware but performs poorly with hidden or unknown malware. Code obfuscation techniques, such as packers, polymorphism and metamorphism, are able to fool current detection techniques by modifying the parent code to produce offspring copies resulting in malware that has the same functionality, but with a different structure. These evasion techniques exploit the drawbacks of traditional malware detection methods, which take current malware structure and create a signature for detecting this malware in the future. However, obfuscation techniques aim to reduce vulnerability to any kind of static analysis to the determent of any reverse engineering process. Furthermore, malware can be hidden in file system slack space, inherent in NTFS file system based partitions, resulting in malware detection that even more difficult.
Description:
Every IPv6-enabled device connected and communicating over the Internet forms the Internet of things (IoT) that is prevalent in society and is used in daily life. This IoT platform will quickly grow to be populated with billions or more objects by making every electrical appliance, car, and even items of furniture smart and connected. The 5th generation (5G) and beyond networks will further boost these IoT systems. The massive utilization of these systems over gigabits per second generates numerous issues. Owing to the huge complexity in large-scale deployment of IoT, data privacy and security are the most prominent challenges, especially for critical applications such as Industry 4.0, e-healthcare, and military. Threat agents persistently strive to find new vulnerabilities and exploit them. Therefore, including promising security measures to support the running systems, not to harm or collapse them, is essential. Nature-inspired algorithms have the capability to provide autonomous and sustainable defense and healing mechanisms. This paper first surveys the 5G network layer security for IoT applications and lists the network layer security vulnerabilities and requirements in wireless sensor networks, IoT, and 5G-enabled IoT. Second, a detailed literature review is conducted with the current network layer security methods and the bio-inspired techniques for IoT applications exchanging data packets over 5G. Finally, the bio-inspired algorithms are analyzed in the context of providing a secure network layer for IoT applications connected over 5G and beyond networks.
Description:
Every IPv6-enabled device connected and communicating over the Internet forms the Internet of things (IoT) that is prevalent in society and is used in daily life. This IoT platform will quickly grow to be populated with billions or more objects by making every electrical appliance, car, and even items of furniture smart and connected. The 5th generation (5G) and beyond networks will further boost these IoT systems. The massive utilization of these systems over gigabits per second generates numerous issues. Owing to the huge complexity in large-scale deployment of IoT, data privacy and security are the most prominent challenges, especially for critical applications such as Industry 4.0, e-healthcare, and military. Threat agents persistently strive to find new vulnerabilities and exploit them. Therefore, including promising security measures to support the running systems, not to harm or collapse them, is essential. Nature-inspired algorithms have the capability to provide autonomous and sustainable defense and healing mechanisms. This paper first surveys the 5G network layer security for IoT applications and lists the network layer security vulnerabilities and requirements in wireless sensor networks, IoT, and 5G-enabled IoT. Second, a detailed literature review is conducted with the current network layer security methods and the bio-inspired techniques for IoT applications exchanging data packets over 5G. Finally, the bio-inspired algorithms are analyzed in the context of providing a secure network layer for IoT applications connected over 5G and beyond networks.