No description available.

No description available.

The image segmentation task divides an image into regions of similar pixels based on brightness, color, and texture, in which every pixel in the image is as- signed to a label. Segmentation is vital in numerous medical imaging applications, such as quantifying the size of tissues, the localization of diseases, treatment plan- ning, and surgery guidance. This thesis focuses on two medical image segmentation tasks: retinal vessel segmentation in fundus images and brain segmentation in 3D MRI images. Finally, we introduce LEON, a lightweight neural network for edge detection. The first part of this thesis proposes a lightweight neural network for retinal blood vessel segmentation. Our model achieves cutting-edge outcomes with fewer parameters. We obtained the most outstanding performance results on CHASEDB1 and DRIVE datasets with an F1 measure of 0.8351 and 0.8242, respectively. Our model has few parameters (0.34 million) compared to other networks such as ladder net with 1.5 million parameters and DCU-net with 1 million parameters. The second part of this thesis investigates the association between whole and re- gional volumetric alterations with increasing age in a large group of healthy subjects (n=6739, age range: 30–80). We used a deep learning model for brain segmentation for volumetric analysis to extract quantified whole and regional brain volumes in 95 classes. Segmentation methods are called edge or boundary-based methods based on finding abrupt changes and discontinuities in the intensity value. The third part of the thesis introduces a new Lightweight Edge Detection Network (LEON). The proposed approach is designed to integrate the advantages of the deformable unit and DepthWise Separable convolutions architecture to create a lightweight back- bone employed for efficient feature extraction. Our experiments on BSDS500 and NYUDv2 show that LEON, while requiring only 500000 parameters, outperforms the current lightweight edge detectors without using pre-trained weights. / Graduate / 2022-10-12

<p>This thesis presents the proof-of-concept of electro-erosion edge (EE) honing as a novel edge preparation process that is based on micro-shaping of the cutting edges of metal cutting tools by electro discharge machining (EDM). This process in its simplest form is first applied to straight edge high speed steel cutting tools which results in a four-fold enhancement in the lives of these tools as compared to the sharp unprepared ones. In the next step the EE-honing application is expanded to hone carbide tools of a complex geometry through the innovative idea of using foil counterfaces. Foil counterface ensures the uniform processing of the entire edge length irrespective of macrogeometric complexities such as curvilinear cutting edges and nose radii. By employing this technique, cutting tools of a complex geometry can be prepared with only 13% edge radius variation which is significantly lower than 40% variation reported for conventional edge preparation processes. ED-machining of cemented carbides necessitates the systematic identification of optimal process parameters to preserve process stability and surface integrity. It is shown that by the application of optimal EE-honing process parameters, EE-honed tools achieved the same life of conventionally prepared ones with the same radii. The advent of advanced edge preparation techniques like EE-honing process has led to the emergence of engineered cutting edge microgeometries most of which cannot be represented by a single edge radius value. In this regard, a novel idea of using parametric quadratic curves for comprehensive cutting edge characterization is presented in the next part of this thesis. The free-knot B-spline approximation enables the unique identification of cutting edge separation points from the clearance and rake faces. Subsequent to the edge identification, quadratic parametric polynomials are employed to characterize the cutting edge by four characterization parameters. These parameters are contour-based and easy to visualize. As the final part of the thesis, the EE-honing process is simulated numerically to gain a better insight into the process. Simulation can model the generation of symmetrical and asymmetrical edges and predict edge geometry with the maximum of 14% error.</p> / Doctor of Philosophy (PhD)

One of the fundamental challenges in Internet-of-Things systems is that network environment is always changing. Conventional networking approaches do not consider the dynamic evaluation of the networks or consider the network dynamic as a mirror thing, which may not be able to work or has a low efficiency in the Internet-of-Things systems. This dissertation is uniquely built by considering the dynamic network environment and even taking advantage of the network dynamic to improve the network performances, with a focus on the routing and offloading issues. The first part is related to the routing design in the opportunistic mobile networks. The opportunistic mobile network is expected to be an intrinsic part of the Internet of Things. Devices communicate with each other autonomously without any centralized control and collaborate to gather, share, and forward information in a multi-hop manner. The main challenge in opportunistic mobile networks is due to intermittent connection and thus data is delivered through store-carry-forwarding paradigm. In this dissertation, We found an observation regarding the contact duration and proposed efficient data partitioning routing algorithms in the opportunistic mobile networks. The second part is related to the offloading issues in the Internet-of-things systems. With the surging demand on high-quality mobile services at any time, from anywhere, how to accommodate the explosive growth of traffics with/without existing network infrastructures is a fundamental issue. Specifically, We consider three different offloading problems, i.e., cellular data offloading, cloud task offloading, and mobile worker task offloading problems in vehicular networks, cloud, and crowdsourcing platforms. The common issue behind them is how to efficiently utilize the network resources in different scenarios by design efficient scheduling mechanisms. For the cellular data offloading, We explored the trade-off of cellular offloading in the vehicular network. For the cloud task offloading, We conducted the research to adjust the offloading strategies wisely so that the total offloading cost is minimized. For the worker task offloading in the smart cities, We optimized the cost-efficiency of the crowdsourcing platforms. / Computer and Information Science

Two major problems in viscoelastic turbulence, the effects of polymers on the laminar-turbulent transition dynamics and the origin of the maximum drag reduction asymptote, can be both better understood in the regime near the margin of turbulence. In the first part of this thesis, direct numerical simulation trajectories initiated from the edge state are used to follow its unstable manifold into the turbulent basin. In Newtonian flow, the growth of turbulence starts with the intensification of velocity streaks and a sharp rise in the Reynolds shear stress. It is followed by a quick breakdown into high-intensity small-scale fluctuations before entering the core of turbulence. Adding drag-reducing polymers does not affect the initial growth of turbulence but stabilizes the primary streak-vortex structure, which help the flow circumvent the breakdown stage. Throughout the process, polymers act in reaction to the growing turbulence and do not drive the instability. This part not only reveals the transition dynamics into turbulence but also presents a comprehensive view of the bursting stage observed in the near-wall self-sustaining cycle, which starts as the flow leaves hibernating turbulence and is redirected towards the turbulent basin by the unstable manifold of the edge state. On the other hand, this thesis also discusses the effects of polymer addition on the laminar-turbulent transition in extended domains. Localized turbulent spot can be clearly observed in the large box, and this turbulent region will spread as well as tend to “split” but finally fill up the whole domain before it is separated. Polymers do not affect the flow dynamics until the burst. Similarly, vortex structures rapidly break down into small scales after the first bursting of Reynolds shear stress, but polymer additives depress this process. The thesis offers a clear and comprehensive overview of the transition into turbulence in the presence of drag-reducing polymers. Future work remains in two major directions. The first is to pinpoint the flow states responsible for the quantitative origin of the universal upper limit of drag reduction observed in experiments. The second is to determine the role, if any, of elasticity-driven instabilities in the transition. / Thesis / Master of Applied Science (MASc) / Turbulence exists everywhere and can be observed in most fluid flows occurring in nature. To reduce the energy consumption, frictional resistance in the turbulence must be considered in fluid transportation. It has been known since the 1940s that a small amount of long-chain polymer additives can dramatically reduce such drag. The mechanism of drag reduction has attracted extensive attention. Two problems of particular interest are the upper limit of drag reduction (termed maximum drag reduction) and the polymer effects on the laminar-turbulent transition. In this thesis, full transient trajectories from marginal turbulent states towards sustained turbulence in both Newtonian and polymeric flows are monitored by direct numerical simulations. It is observed that polymer additives do not affect the initial growth of turbulence but prevent flows from breaking into strong but small-scale fluctuations afterwards. In a more extended domain, turbulence starts as localized spots which spread across the channel. Adding polymers changes the dynamics of turbulence propagation as well. In addition to the aforementioned problems, this study also sheds lights on the so-called bursting events intermittent surges in turbulent activities observed in experiments.

We study the spontaneous edge current of chiral superconductors with high chirality both in the absence and presence of Meissner screening. We compute the edge current from a self-consistent solution to a set of coupled equations: quasiclassical Eilenberger equation, superconducting gap equation, and Maxwell equation. We find that the spatial dependent chiral edge current is largely suppressed and has more nodes for higher chirality pairings. In the absence of Meissner screening, the integrated current at T=0 is zero for all higher chirality pairings; while it is substantial for chiral p-wave. This conclusion is consistent with previous studies. In contrast, at finite T, the integrated current is non-zero even for higher chiral pairings. It turns out that the spatial varying order parameter is crucial to understand this finite T behavior of the edge current. When Meissner screening is included, the magnitude of the edge currents is reduced for all chiral pairings; however, the reduction is much weaker in higher chirality cases. We conclude that the Meissner effect is not that important for higher chiral pairings. We also consider the effect of the rough surface on the edge current. The edge current of even chiral pairings is inverted by the strong surface roughness; however, that of the odd chiral pairings is not. The sub-dominant order parameters, induced by the surface, are the key to understanding this current inversion. / Thesis / Master of Science (MSc)

Cutting tool geometries play important roles in tool performance, such as tool life, surface integrity, and cutting force. The most common commercial tools edge geometries are honed, chamfered and hone-chamfered. This study investigates new ways to develop tool geometry. An uncoated carbide tool is used in the orthogonal cutting of AISI 4140. By observing the tool geometry changes in the machining process with white light interferometry, a new tool wear geometry model has been proposed. A non-destructive tool wear measurement is discussed as well. In addition, this study presents the machining result comparison between the new and conventional geometries as well as the failure analysis from both experimental and FEA perspectives. / Thesis / Master of Applied Science (MASc)

The limited energy and computational resources in small-scale smart devices impede the expansion of the range of applications that those devices can support, especially to applications with tight latency constraints. Mobile edge computing is a promising framework that provides shared computational resources in the access points in the network and provides devices in that network with the opportunity to offload (a portion of) their computational tasks to the access points. To effectively capture that opportunity in an offloading system with multiple devices, the available communication and computation resources must be efficiently allocated. The main focus of this thesis is on the optimal allocation of communication resources in a K-user offloading system. The resource allocation problem that is considered in this thesis captures minimizing the total energy consumption of users while the requirements of the users, and their computational tasks, are met. That problem is addressed for two of the most widely-considered classes of computational tasks in the literature, namely, indivisible tasks (binary offloading) and divisible tasks (partial offloading). This thesis begins with an exploration of the impact of the choice of multiple access scheme that is employed by the system on the total energy consumption of the users. In particular, the problem of minimizing the total energy consumption of a two-user binary offloading system is tackled under various multiple access schemes, namely time division multiple access (TDMA), sequential decoding without time sharing, independent decoding, and multiple access schemes that can exploit the full capabilities of the channel, which are referred to as full multiple access schemes (FullMA) in this thesis. Using a decomposition-based approach, closed-form solutions to the resource allocation problem are obtained. Those expressions show that by exploiting the full capabilities of the channel, a FullMA scheme can significantly reduce the total energy consumption of the users as compared to the other schemes. The closed-form expressions also show that when the channel gains of the two users are equal, the TDMA scheme can achieve the optimal energy consumption. For the case of partial offloading, an analogous analysis leads to a reduced-dimension design problem and an extension to the optimally result for TDMA. In the next step of the development, the insights obtained from the decomposition-based analysis of the two-user case are used to tackle the communication resource allocation problem for a K-user offloading system in which the users are assumed to be served over a single time slot. Based on their performance in the two-user case, FullMA and TDMA schemes are considered. The mixed-integer optimization problem that arises in the binary offloading case is addressed by employing a decomposition approach with a closed-form expression obtained for the optimal resource allocation for given offloading decisions, and a tailored pruned greedy search algorithm developed herein for the offloading decisions. By exploiting the maximum allowable latency of each individual user, the proposed algorithm is able to significantly reduce the energy consumption of the users in comparison to the existing algorithms in the literature that assume equal latency constraints for all users. Furthermore, with the closed-form optimal solution to the resource allocation problem obtained for given offloading decisions, the proposed algorithm has a significantly lower computational cost compared to the existing algorithms. In the partial offloading case, a quasi-closed- form solution is obtained for the resource allocation problem. Finally, a time-slotted signalling structure is proposed as an optimal transmission structure for a generic K-user offloading system. Furthermore, an optimal times-lotted structure that requires only K time slots is developed for a K-user offloading system that employs a FullMA scheme. The proposed time-slotted structure not only exploits the maximum latency constraint of each user, it also exploits the differences between the latency constraints of the users by taking advantage of the interference reduction that arises when a user finishes offloading. The proposed time-slotted FullMA signalling structure significantly reduces the energy consumption of the users compared to some existing methods that employ the TDMA scheme, and compared to those with FullMA, but sub-optimal single-time-slot signalling structures. Moreover, the computational cost of the proposed time-slotted algorithm is significantly lower than that of the existing algorithms in the literature. / Dissertation / Doctor of Philosophy (PhD) / The rapid increase in the number of smart devices in wireless communication networks, and the expansion in the range of computationally-intensive and latency sensitive applications that those devices are required to support, have highlighted their resource limitations in terms of energy, power, central processing unit (CPU), and memory. Mobile edge computing is a framework that provides shared computational resources at the access points of wireless networks and gives such devices the opportunity to offload (a portion of) their applications to be executed at the access points. In order to fully exploit such an opportunity when multiple devices seek to offload their applications, the available communication and computation resources must be efficiently allocated amongst those devices. The ultimate goal of this thesis is to obtain the optimal communication resource allocation in a K-user offloading system while different constraints on the devices and on the applications are satis ed. To that end, this thesis shows that the minimum energy consumption is obtained when the system exploits the full capabilities of the channel, the maximum allowable latency of each user, and the differences between the latency constraints of each user. Accordingly, this thesis proposes an optimized signalling structure and, based on that structure, low-complexity algorithms that achieve an energy-optimal resource allocation in a K-user offloading system.

MACHINABILITY ENHANCEMENT OF STAINLESS STEELS THROUGH CONTROL OF BUILT-UP EDGE FORMATION / Demand for parts made from stainless steel is rapidly increasing, especially in the oil and gas industries. Stainless steel provides a number of key advantages, such as high tensile strength, toughness, and excellent corrosion resistance. However, stainless steel cutting faces some serious difficulties. At low cutting speeds, workpiece material and the chips formed during machining tend to adhere to the cutting tool surface, forming a built-up edge (BUE). The BUE is an extremely deformed piece of material which intermittently sticks to the tool at the tool-chip interface throughout the cutting test, affecting tool life and surface integrity. Unstable BUE can cause tool failure and deterioration of the workpiece. However, stable BUE formation can protect the cutting tool from further wear, improving the productivity of stainless steel machining. This thesis presents an in-depth study of machining performance using different coated tools and various coolant conditions to examine the nature of the different tool wear mechanisms present during the turning of stainless steels. Then, different textures are generated on the tool rake face to control the stability of BUE and reduce friction during the machining process. Results show that the BUE can significantly improve the frictional conditions and workpiece surface integrity at low cutting speeds. Finally, square textures on tool rake face were found to control the stability of BUE and minimize the friction at the tool-chip interface. This reduces the average coefficient of friction by 20-24% and flank wear by 41-78% and increases surface finish by 54-68% in comparison to an untextured tool. / Thesis / Doctor of Philosophy (PhD) / Three main objectives are presented in this thesis. The first is a detailed investigation of the performance of stainless steel machining obtained by the use of different coated cutting tools and various cooling conditions. The goal of this research is to assess the reduction of tool service life, productivity, and part quality. The thesis also examines the causes of workpiece material adhesion to the cutting tool during the cutting test and to better explain its effects on tool wear and workpiece surface finish. This phenomenon is known as the "built-up edge" (BUE). Finally, different textures are applied on the cutting tool via a laser to stabilize the BUE formation on the cutting tool, thereby improving the quality of the part.