Computing Journal Articles

Permanent URI for this collection


Recent Additions

Now showing 1 - 5 of 54
  • Item
    Feature-based systematic analysis of advanced persistent threats
    (InTech, 2023-05-22) Miguez, M.; Sarrafpour, Bahman; Unitec, Te Pūkenga; Te Pūkenga
    Advanced Persistent Threats (APT) and Targeted Attacks (TA) targeting high-value organizations continue to become more common. These slow (sometimes carried on over the years), fragmented, distributed, seemingly unrelated, very sophisticated, highly adaptable, and, above all, stealthy attacks have existed since the large-scale popularization of computing in the 1990s and have intensified during the 2000s. The aim of attackers has expanded from espionage to attaining financial gain, creating disruption, and hacktivism. These activities have a negative impact on the targets, many times costing significant amounts of money and destabilizing organizations and governments. The resounding goal of this research is to analyze previous academic and industrial research of 72 major APT attacks between 2008 and 2018, using 12 features, and propose a categorization based on the targeted platform, the time elapsed to discovery, targets, type, purpose, propagation methods, and derivative attacks. This categorization provides a view of the effort of the attackers. It aims to help focus the design of intelligent detection systems on increasing the percentage of discovered and stopped attacks.
  • Item
    Analysis and comparison of deepfakes detection methods for cross-library generalisation
    (2023-08-21) Wang, Changjin; Sharifzadeh, Hamid; Varastehpour, Soheil; Ardekani, Iman; Unitec, Te Pūkenga; Te Pūkenga
    The rise of generative artificial intelligence (GenAI) has made it increasingly possible to use Deepfakes technology to generate fake pictures and videos. While this technology has benefits, it also has downsides such as spreading misinformation and endangering public interests. To address this issue, researchers have proposed various deep forgery detection algorithms and have achieved remarkable results. However, a common problem regarding these detection methods is that while in-library detection can usually achieve high accuracy, their performance is significantly degraded in cross-library detection. This indicates a severe problem of insufficient generalisation ability. To better compare the performance differences between various detection methods, this paper analyses the detection performance of the six established models of Two-stream, MesoNet, HeadPose, FWA, VA, and Multi-task. To ensure consistency, we employ a uniform evaluation framework as a benchmark for comparison. We conduct extensive intra-library and crosslibrary tests to evaluate these methods’ generalisation ability by utilising accuracy and error rate as key evaluation criteria for our experiments. Additionally, we further explore areas for improvement by analysing the impact of data augmentation, dataset partitioning, and threshold selection on the performance of these detection methods. Our comparative experiments are conducted on three existing fake face video datasets, including FaceForensics++, DeepfakeTIMIT, and Celeb-DF. Our research findings indicate the database partitioning method has a direct impact on the detector’s performance, and to enhance generalisation performance, the database should be divided person-based manually. The effectiveness of data augmentation techniques in improving cross-library performance is generally limited, and setting the threshold directly using source domain data often leads to a high error rate in the target domain. The findings of this paper provide insights into the development of more effective detection methods to combat the harmful effects
  • Item
    Critical data detection for dynamically adjustable product quality in IIoT-enabled manufacturing
    (Institute of Electrical and Electronics Engineers (IEEE), 2023-05-17) Sen, Sachin K.; Karmakar, G. C.; Pang, Shaoning; Unitec,Te Pūkenga; Te Pūkenga; Institute of Innovation, Science and Sustainability (Federation University Australia)
    The IIoT technologies, due to the widespread use of sensors, generate massive data that are key in providing innovative and efficient industrial management, operation, and product quality control processes. The significance of data has prompted relevant research communities and application developers how to harness the values of these data in secure manufacturing. Critical data analysis, identification of critical factors to improve the manufacturing process and critical data associated with product quality have been investigated in the current literature. However, the current works on product quality control are mainly based on static data analysis, where data may change, but there is no way to adjust them dynamically. Thus, they are not applicable for product quality control, at which point their adjustment is instantly required. However, many manufacturing systems exist, like beverages and food, where ingredients must be adjusted instantaneously to maintain product quality. To address this research gap, we introduce a method that identifies the critical data based on their ranking by exploiting three criticality assessment criteria that capture the instantaneous product quality change during manufacturing. These three criteria are – (1) correlation, (2) percentage quality change and (3) sensitivity for the assessment of data criticality. The product quality is estimated using polynomial regression (POLY), SVM, and DNN. The proposed method is validated using wine manufacturing data. Our proposed method accurately identifies critical data, where SVM produces the lowest average production quality prediction error (10.40%) compared with that of POLY (11%) and DNN (14.40%) This work was supported by the Defence Science Institute (DSI), Australia, through the DSI-Facilitated Project
  • Item
    Using vector agents to implement an unsupervised image classification algorithm
    (MDPI (Multidisciplinary Digital Publishing Institute), 2021-12-02) Borna, Kambiz; Moore, T.; Azadeh, N.H.; Pascal, S.; Unitec Institute of Technology; University of Otago; Federation University Australia
    Unsupervised image classification methods conventionally use the spatial information of pixels to reduce the effect of speckled noise in the classified map. To extract this spatial information, they employ a predefined geometry, i.e., a fixed-size window or segmentation map. However, this coding of geometry lacks the necessary complexity to accurately reflect the spatial connectivity within objects in a scene. Additionally, there is no unique mathematical formula to determine the shape and scale applied to the geometry, being parameters that are usually estimated by expert users. In this paper, a novel geometry-led approach using Vector Agents (VAs) is proposed to address the above drawbacks in unsupervised classification algorithms. Our proposed method has two primary steps: (1) creating reliable training samples and (2) constructing the VA model. In the first step, the method applies the statistical information of a classified image by k-means to select a set of reliable training samples. Then, in the second step, the VAs are trained and constructed to classify the image. The model is tested for classification on three high spatial resolution images. The results show the enhanced capability of the VA model to reduce noise in images that have complex features, e.g., streets, buildings.
  • Item
    An approach to interesting objects detection in low quality image sequences for fisheries management
    (World Scientific and Engineering Academy and Society (WSEAS), 2015) Li, Z.; Chen, L.; Peng, J.; Song, Lei; Unitec Institute of Technology; Chongqing University of Science and Technology
    In order to extract interest objects in low quality image sequences from fisheries management, this paper proposes a new significant feature extraction method based on cascade framework. This algorithm involves pre processing image sequences, clipping interesting areas, extracting SURF features, removing boundary features, and acquiring significant features with interesting objects. We apply our algorithm to fisheries management for counting and matching ships and cars, the proposed method can efficiently detect multiple objects from real-scene video frames with averaged accuracy 91.63%