No description available.

The animal-vehicle collision has been a topic of concern for years, especially in North America. To mitigate the problem, this thesis focuses on animal detection based on the onboard camera for intelligent vehicles. In the domain of image classification and object detection, the methods of shape matching and local feature crafting have reached the technical plateau for decades. The development of Convolutional Neural Network (CNN) brings a new breakthrough. The evolution of CNN architectures has dramatically improved the performance of image classification. Effective frameworks on object detection through CNN structures are thus boosted. Notably, the family of Region-based Convolutional Neural Networks (R-CNN) perform well by combining region proposal with CNN. In this thesis, we propose to apply a new region proposal method|Maximally Stable Extremal Regions (MSER) in Fast R-CNN to construct the animal detection framework. MSER algorithm detects stable regions which are invariant to scale, rotation and viewpoint changes. We generate regions of interest by dealing with the result of MSER algorithm in two ways: by enclosing all the pixels from the resulted pixel-list with a minimum enclosing rectangle (the PL MSER) and by fitting the resulted elliptical region to an approximate box (the EL MSER). We then preprocess the bounding boxes of PL MSER and EL MSER to improve the recall of detection. The preprocessing steps consist of filtering out undesirable regions by aspect ratio model, clustering bounding boxes to merge the overlapping regions, modifying and then enlarging the regions to cover the entire animal. We evaluate the two region proposal methods by the measurement of recall over IOU-threshold curve. The proposed MSER method can cover the expected regions better than Edge Boxes and Region Proposal Network (RPN) in Faster R-CNN. We apply the MSER region proposal method to the framework of R-CNN and Fast R-CNN. The experiments on the animal database with moose, deer, elk, and horses show that Fast R-CNN with MSER achieves better accuracy and faster speed than R-CNN with MSER. Concerning the two ways of MSER, the experimental results show that PL MSER is faster than EL MSER and EL MSER gains higher precision than PL MSER. Also, by altering the structure of network used in Fast R-CNN, we verify that network stacking more layers achieves higher accuracy and recall. In addition, we compare the Fast R-CNN framework using MSER region proposal with the state-of-the-art Faster R-CNN by evaluating the experimental results of on our animal database. Using the same CNN structure, the proposed Fast R-CNN with MSER gains a higher average accuracy of the animal detection 0.73, compared to 0.42 of Faster R-CNN. In terms of detection quality, the proposed Fast R-CNN with MSER achieves better IoU histogram than that of Faster R-CNN.

L'auteur n'a pas fourni de résumé en français / L'auteur n'a pas fourni de résumé en anglais

Doctor of Philosophy / Department of Mechanical and Nuclear Engineering / Douglas S. McGregor / Interest in high-efficiency replacements for thin-film-coated thermal neutron detectors led to the development of single-sided microstructured semiconductor neutron detectors (MSNDs). MSNDs are designed with micro-sized trench structures that are etched into a vertically-oriented pvn-junction diode, and backfilled with a neutron converting material, such as ⁶LiF. Neutrons absorbed by the converting material produce a pair of charged-particle reaction products that can be measured by the diode substrate. MSNDs have higher neutron-absorption and reaction-product counting efficiencies than their thin-film-coated counterparts, resulting in up to a 10x increase in intrinsic thermal neutron detection efficiency. The detection efficiency for a single-sided MSND is reduced by neutron streaming paths between the conversion-material filled regions that consequently allow neutrons to pass undetected through the detector. Previously, the highest reported intrinsic thermal neutron detection efficiency for a single MSND was approximately 30%. Methods for double-stacking and aligning MSNDs to reduce neutron streaming produced devices with an intrinsic thermal neutron detection efficiency of 42%. Presented here is a new type of MSND that features a complementary second set of trenches that are etched into the back-side of the detector substrate. These dual-sided microstructured semiconductor neutron detectors (DS-MSNDs) have the ability to absorb and detect neutrons that stream through the front-side, effectively doubling the detection efficiency of a single-sided device. DS-MSND sensors are theoretically capable of achieving greater than 80% intrinsic thermal neutron detection efficiency for a 1-mm thick device. Prototype DS-MSNDs with diffused pvp-junction operated at 0-V applied bias have achieved 53.54±0.61%, exceeding that of the single-sided MSNDs and double-stacked MSNDs to represent a new record for detection efficiency for such solid-state devices.

Indiana University-Purdue University Indianapolis (IUPUI) / Trained canines are commonly used as biological detectors for explosives; however, there are some areas of uncertainty that have led to difficulties in canine training and testing. Even though a standardized container for determining the accuracy of explosives-detecting canines has already been developed, the factors that govern the amount of explosive vapor that is present in the system are often uncertain. This has led to difficulties in comparing the sensitivity of canines to one another as well as to analytical instrumentation, despite the fact that this container has a defined headspace and degree of confinement of the explosive. For example, it is a common misconception that the amount of explosive itself is the chief contributor to the amount of odor available to a canine. In fact, odor availability depends not only on the amount of explosive material, but also the explosive vapor pressure, the rate with which the explosive vapor is transported from its source and the degree to which the explosive is confined. In order to better understand odor availability, headspace GC/MS and mass loss experiments were conducted and the results were compared to the Ideal Gas Law and Fick’s Laws of Diffusion. Overall, these findings provide increased awareness about availability of explosive odors and the factors that affect their generation; thus, improving the training of canines. Another area of uncertainty deals with the complexity of the odor generated by the explosive, as the headspace may consist of multiple chemical compounds due to the extent of explosive degradation into more (or less) volatile substances, solvents, and plasticizers. Headspace (HS) and solid phase microextraction (SPME) coupled with gas chromatography/mass spectrometry (GC/MS) were used to determine what chemical compounds are contained within the headspace of an explosive as well as NESTT (Non-Hazardous Explosive for Security Training and Testing) products. This analysis concluded that degradation products, plasticizers, and taggants are more common than their parent explosive.

This master thesis focuses on the cutting-edge application of AI in developing intrusion detection systems (IDS) for unmanned aerial vehicles (UAVs) in smart cities. The objective is to address the escalating problem of UAV intrusions, which pose a significant risk to the safety and security of citizens and critical infrastructure. The thesis explores the current state of the art and provides a comprehensive understanding of recent advancements in the field, encompassing both physical and network attacks. The literature review examines various techniques and approaches employed in the development of AI-based IDS. This includes the utilization of machine learning algorithms, computer vision technologies, and edge computing. A proposed solution leveraging computer vision technologies is presented to detect and identify intruding UAVs in the sky effectively. The system employs machine learning algorithms to analyze video feeds from city-installed cameras, enabling real-time identification of potential intrusions. The proposed approach encompasses the detection of unauthorized drones, dangerous UAVs, and UAVs carrying suspicious payloads. Moreover, the thesis introduces a Cycle GAN network for image denoising that can translate noisy images to clean images without the need for paired training data. This approach employs two generators and two discriminators, incorporating a cycle consistency loss that ensures the generated images align with their corresponding input images. Furthermore, a distributed architecture is proposed for processing collected images using an edge-offloading approach within the UAV network. This architecture allows flying and ground cameras to leverage the computational capabilities of their IoT peers to process captured images. A hybrid neural network is developed to predict, based on input tasks, the potential edge computers capable of real-time processing. The edge-offloading approach reduces the computational burden on the centralized system and facilitates real-time analysis of network traffic, offering an efficient solution. In conclusion, the research outcomes of this thesis provide valuable insights into the development of secure and efficient IDS for UAVs in smart cities. The proposed solution contributes to the advancement of the UAV industry and enhances the safety and security of citizens and critical infrastructure within smart cities.

Data drifts naturally occur in data streams due to seasonality, change in data usage, and the data generation process. Concepts modelled via the data streams will also experience such drift. The problem of differentiating concept drift from anomalies is important to identify normal vs abnormal behaviour. Existing techniques achieve poor responsiveness and accuracy towards this differentiation task. We take two approaches to address this problem. First, we extend an existing sliding window algorithm to include multiple windows to model recently seen data stream patterns, and define new parameters to compare the data streams. Second, we study a set of optimisers and tune a Bi-LSTM model parameters to maximize accuracy. / Thesis / Master of Applied Science (MASc)

No description available.

Program attacks are one of the oldest and fundamental cyber threats. They compromise the confidentiality of data, the integrity of program logic, and the availability of services. This threat becomes even severer when followed by other malicious activities such as data exfiltration. The integration of primitive attacks constructs comprehensive attack vectors and forms advanced persistent threats. Along with the rapid development of defense mechanisms, program attacks and data leak threats survive and evolve. Stealthy program attacks can hide in long execution paths to avoid being detected. Sensitive data transformations weaken existing leak detection mechanisms. New adversaries, e.g., semi-honest service provider, emerge and form threats. This thesis presents theoretical analysis and practical detection mechanisms against stealthy program attacks and data leaks. The thesis presents a unified framework for understanding different branches of program anomaly detection and sheds light on possible future program anomaly detection directions. The thesis investigates modern stealthy program attacks hidden in long program executions and develops a program anomaly detection approach with data mining techniques to reveal the attacks. The thesis advances network-based data leak detection mechanisms by relaxing strong requirements in existing methods. The thesis presents practical solutions to outsource data leak detection procedures to semi-honest third parties and identify noisy or transformed data leaks in network traffic. / Ph. D.

This study investigates the real life feasibility of applying modern estimation theory to target angular measurement information provided by a short to medium range, lightweight, tactical, tracking radar. Techniques are considered in terms of their computational demand and their effectiveness in filtering practically obtained measurement data. With the aid of a mathematical model, the angular measurement operation of the radar is shown to be unlikely to provide the desired measurement information for these operating conditions. This deficiency is due to multisource noise and encompasses such well known phenomena as glint. Analysis of measurement data obtained from T.V. and radar trials conducted using the Marconi ST802 radar to track a light aircraft, demonstrates these phenomena. Standard Kalman solutions proposed in the literature are applied to these measurements and shown to be ineffective against multisource noise. Consequently modifications are proposed and shown to be considerably more effective. The Success of these modifications led to their application to a low elevation angle tracking example, where multisource noise can severely degrade the performance of the radar. As a result, further tests with low angle data are recommended.