Browsing by Author "Alim, Usman"
Now showing 1 - 20 of 40
Results Per Page
Sort Options
- ItemOpen Access3De Interactive Lenses for Visualization in Virtual Environments(2018-10) Mota, Roberta Cabral Ramos; Rocha, Allan; Silva, Julio Daniel; Alim, Usman; Sharlin, EhudWe present 3De lens, a technique for focus+context 3D visualization of multiple geometric representations. Our lens fuses two categories of lenses (3D and Decal) into a single coherent entity, thus enabling flexible use of either one or the two lenses combined depending on the underlying data geometry. In addition, we incorporate our lens into virtual reality as it enables a rich and natural style of direct spatial manipulation for exploratory 3D data analysis. To demonstrate its potential use, we discuss two domain examples in which our lens technique creates customized visualizations of both surfaces and streamlines.
- ItemOpen AccessA Deep Learning Based Method for Fast Registration of Cardiac Magnetic Resonance Images(2024-04-29) Graham, Benjamin; Jacob, Christian; Alim, Usman; Zhao, Richard; Garcia Flores, Julio; Hoyer, PeterImage registration is used in many medical image analysis applications, such as tracking the motion of tissue in cardiac images, where cardiac kinematics can be an indicator of tissue health. Registration is a challenging problem for deep learning algorithms because ground truth transformations are not feasible to create, and because there are potentially multiple transformations that can produce images that appear correlated with the goal. Unsupervised methods have been proposed to learn to predict effective transformations, but these methods take significantly longer to predict than established baseline methods. For a deep learning method to see adoption in wider research and clinical settings, it should be designed to run in a reasonable time on common, mid-level hardware. Fast methods have been proposed for the task of image registration but often use patch-based methods which can affect registration accuracy for a highly dynamic organ such as the heart. In this thesis, a fast, volumetric registration model is proposed for the use of quantifying cardiac strain. The proposed Deep Learning Neural Network (DLNN) is designed to utilize an architecture that can com-pute convolutions incredibly efficiently, allowing the model to achieve registration fidelity similar to other state-of-the-art models while taking a fraction of the time to perform inference. The proposed fast and lightweight registration (FLIR) model is used to predict tissue motion which is then used to quantify the non-uniform strain experienced by the tissue. For acquisitions taken from the same patient at approximately the same time, it would be expected that strain values measured between the acquisitions would have very small differences. Using this metric, strain values computed using the FLIR method are shown to be very consistent.
- ItemOpen AccessA Demographic-driven Peer-Assisted CDN for Sharing User-Generated Media Content(2017) Rougeau, Bradley; Wang, Mea; Tang, Tony; Alim, Usman; Wang, MeaInternet traffic is increasingly dominated by user-generated content, predominantly by multimedia content (photos and videos). The content is primarily shared in online social networks (OSNs) such as Pinterest, Twitter, and Facebook. In this thesis, we are interested in studying the traffic imposed by user-generated multimedia content in OSNs and exploring the potential of involving end-user devices for more efficient delivery of user-generated media content. To do so, we developed Viewcount, a Facebook application, to collect traces of user-generated multimedia traffic. Through analyzing social activities around user-generated multimedia content (such as user demographics and viewing distributions), we drew insightful observations correlating demographics and social activities to network traffic. These observations shed light on the design of DEPA, a demographic-driven peer-assisted content delivery network (CDN) to reduce redundant network traffic and workload on the OSN servers. In contrast to existing attempts towards peer-assisted OSNs, DEPA adapts existing OSN architecture and introduces various demographic-driven caching schemes. Through a trace-driven simulator, we evaluated DEPA in both the Viewcount network and on OSNs with different demographic properties. DEPA provided a significant reduction of traffic on the OSN servers even with very small caches on our peer-assistants. Furthermore, the demographic caching algorithm in DEPA is more adaptive to dynamic changes of user online status in OSNs.
- ItemEmbargoA Non-Invasive Approach to Investigating the Intrinsic Cerebrovascular Regulation and Sympathetic Nervous System Using Functional Optical Coherence Tomography (f-OCT)(2024-05-14) Safarzadeh, Mohammad; Wilson, Richard; Alim, Usman; Gordon, GrantWhile cerebral autoregulation and the sympathetic nervous system are crucial for maintaining stable blood flow to the brain, dysfunctions in these systems can lead to severe disorders, including cardiovascular disease and strokes. To enhance future exploration of these phenomena in humans, this thesis investigates retinal and choroidal vasculature changes under hypoxia using functional Optical Coherence Tomography (f-OCT) imaging. The first hypothesis aims to assess cerebrovascular autoregulation through retinal vasculature monitoring, while the second examines sympathetic nervous system activity by observing choroidal vasculature changes, within the White Mountain database. This investigation seeks to enhance tools for monitoring cerebrovascular function and associated diseases. OCT imaging faces challenges like high noise, artifacts, lack of standardization, and assessment difficulties. To address these, two interconnected vascular bed analysis pipelines were used, along with a novel framework named Q-MCDM (Quasi-Monte Carlo Multi Criteria Decision making). Q-MCDM enhances OCT image processing and evaluation, marking a step towards Interpretable AI in healthcare by optimizing design, execution, and trade-offs in image analysis approaches. Using Q-MCDM for the retinal vascular bed, Frangi's algorithm emerged as the optimal choice for segmentation, achieving a Matthews Correlation Coefficient (MCC) of 0.82, Length (LE) of 0.81, and Area (AR) of 0.86. Conversely, in the choroid vascular bed, the ISODATA method yielded the best results with a Dice coefficient of 0.82, Matthews Correlation Coefficient (MCC) of 0.71, and Area (AR) of 0.79. Significant correlations were observed between vessel perfusion density (VPD) and partial O2 pressure, consistently present across best solutions (r value range = -0.701 to -0.663). Distinct behaviors were observed in the choroid vascular bed's layers. The superficial layer showed increased VPD values, potentially indicating vessel dilation due to parasympathetic or intrinsic autoregulation. In contrast, the deep layer exhibited significant constriction in VPD values, possibly linked to sympathetic nervous system activity. Test/retest assessments confirmed the stability of biological response over time, emphasizing the significance of the findings in understanding intrinsic autoregulation and autonomic nervous system through f-OCT.
- ItemOpen AccessAugmented QArt: Interactive Art Recreation by Mobile Augmented Reality(2024-03-21) Booshehrian, Mahnaz; Alim, Usman; Hushlak, Gerald; Sharlin, EhudThe integration of technology into art galleries and museums has opened new avenues for enhancing visitors' experiences and addressing challenges related to accessibility, interactivity, and artistic attraction. Inspired by Picasso’s Cubism paintings and Augmented Reality technology, this thesis explores the design of aesthetically detectable codes, known as QArt codes, to enhance accessibility and visual appeal and to enrich traditional art experiences in settings where space and time are limited. The project aims to answer questions about how technology can be leveraged to enhance artistic environments, offer interactive access to artwork information, and seamlessly blend the actual and virtual realms. Through the implementation of QArt codes, people can easily access and explore various artworks, including Picasso's Harlequin paintings, while saving time and money. This approach offers visitors a chance to learn about different aspects of the artwork, such as style, techniques, and historical context. Interacting with QArt codes provides an engaging and memorable experience, leaving a lasting impression on visitors. A qualitative user study conducted at the end of the project revealed that QArt codes can provide a nuanced understanding of Picasso's artistic range and insight into the Cubism style through multiple dimensions. Overall, QArt codes are an effective and captivating way to share artworks with the world.
- ItemOpen AccessCINAPACT-Splines: A Family of Infnite Smooth, Accurate and Compactly Supported Splines(2015-09-01) Akram, Bita; Alim, Usman; Samavati, FaramarzWe introduce a class of compactly supported C infinite kernels (CINAPACT-splines) whose integer translates form a shift-invariant reconstruction space that can be tuned to achieve any order of accuracy. CINAPACT-splines resemble traditional B-splines in that higher orders of accuracy are achieved by successive convolutions with a B-spline of degree zero. Unlike B-splines however, the starting point for CINAPACT-splines is a compactly supported bump function that has been properly normalized so that it fulfills the partition of unity criterion. We explore the properties of CINAPACT-splines in reconstructing volumetric data sampled on regular grids. We show that CINAPACT-splines provide similar reconstruction quality and cost compared to some well-established filters, while being infinitely smooth. We further explore the advantages of our filter by implementing a curvature-based transfer function using second derivatives of the filter to demonstrate feature lines of a function. We apply the same technique using filters of smaller support and less cost.
- ItemOpen AccessCompactly Supported Biorthogonal Wavelet Constructions on the A-star Lattice and their Application to Visualization(2017) Horacsek, Joshua; Alim, Usman; Alim, Usman; Samavati, Faramarz; Nielsen, John; Li, ZongpengIn this dissertation, a family of compact biorthogonal wavelet filter banks that are tailored to the geometry of the A-star lattice are derived. Our application of interest is on the three dimensional A-star or {\em body centered cubic} (BCC) lattice. While the BCC lattice has been shown to have superior approximation properties for volumetric data when compared to the Cartesian cubic (CC) lattice, there has been little work in the way of designing wavelet filter banks that respect the geometry of the BCC lattice. Since wavelets have applications in signal de-noising, compression, and sparse signal reconstruction, these filter banks are an important tool that addresses some of the scalability concerns presented by the BCC lattice. We use these filters in the context of volumetric data compression and reconstruction and qualitatively evaluate our results by rendering images of isosurfaces from compressed data.
- ItemOpen AccessComparative Visualizations of Noisy and Filtered Blood Flow from 4D PC-MRI Cardiac Datasets(2017-10) Khan, Fahim Hasan; Rocha, Allan; Alim, UsmanModern phase-contrast magnetic resonance imaging (PC-MRI) can acquire both cardiac anatomy and flow function in a single acquisition and deliver high quality volumetric and time-varying (4D) datasets which enable better diagnosis and risk assessment of various cardiovascular diseases. A good way to visualize blood flow from 4D PC-MRI datasets is to use animated pathlines through the anatomical context for representing the trajectories of the blood particles. Artifact correction is one crucial step in the processing pipeline of 4D PC-MRI datasets for representing the cardiac flow using pathlines, which in turn can reduce the overall quality of the useful information in the dataset. In this work, an approach is presented for comparative visualization of 4D PC-MRI datasets before and after artifact correction for qualitative analysis.
- ItemOpen AccessComponent-wise Interpolation of Solenoidal Vector Fields: A Comparative Numerical Study(2015-07-22) Yin, Dandong; Alim, UsmanVector-field interpolation is a fundamental task in flow simulation and visualization. The common practice is to interpolate the vector field in a component-wise fashion. When the vector field of interest is solenoidal (divergencefree), such an approach is not conservative and gives rise to artificial divergence. In this work, we numerically compare some recently proposed scalar interpolation methods on the Cartesian and body-centered cubic lattices, and investigate their ability to conserve the solenoidal nature of the vector field. We start with a sampled version of a synthetic solenoidal vector field and use an interpolative component-wise reconstruction method to approximate the vector field and its divergence at arbitrary locations. Our results show that an improved scalar interpolation method does not necessarily lead to a more conservative vector field approximation.
- ItemOpen AccessCompressive Volume Rendering(2015-07-09) Liu, Xiaoyang; Alim, UsmanCompressive rendering refers to the process of reconstructing a full image from a small subset of the rendered pixels, thereby expediting the rendering task. Images produced via direct volume rendering are usually highly compressible in a transform domain such as the Fourier or wavelet domains. In this dissertation, we empirically investigate four image order tech- niques for compressive rendering that are suitable for direct volume rendering. The first technique is based on the theory of compressed sensing and leverages the sparsity of the image gradient in the Fourier domain. Following this, we investigate sparse representation of volume rendered images via dictionary learning. The latter techniques exploit smoothness properties of the rendered image; the third technique recovers the missing pixels via a to- tal variation minimization procedure while the fourth technique incorporates a smoothness prior in a variational reconstruction framework employing interpolating cubic B-splines. We compare and contrast these four techniques in terms of quality, efficiency and sensitivity to the distribution of pixels. Our results show that smoothness-based techniques significantly outperform techniques that are based on compressed sensing as well as dictionary learning and are also robust in the presence of highly incomplete information. We achieve high quality recovery with as little as 20% of the pixels distributed uniformly in screen space.
- ItemOpen AccessContinuous and Discrete Data-Processing on Non-Cartesian Lattices(2024-06-21) Horacsek, Joshua; Alim, Usman; Korobenko, Artem; Ware, Anthony; Ioannou, Yani; Entezari, AlirezaThis thesis focuses on the challenges and practical considerations involved in approximating natural phenomena on regular, yet non-square (non-Cartesian) grids. At a high level, the most simple illustrative example is the move away from square pixels, to say, hexagonal pixels, which have much nicer symmetry compared to square pixels. The focus of this work is geared towards pragmatic solutions, building theory when needed, but with careful consideration to the practical aspects of data processing found in many sub-domains of computer science. The key sub-domains we explore are primarily scientific visualization and machine learning; but the techniques within this thesis extend much further into the numerical sciences. Central to our exploration is the development of the lattice tensor. This data structure is de- signed to encapsulate the complexities inherent in handling non-Cartesian grids. The lattice tensor is simple enough in its formulation so as to be integrated within an (auto)-differentiable com- putational framework. This immediately opens machine learning to the world of non-Cartesian methods. In addition to introducing the lattice tensor, this thesis proposes and evaluates various practical methods for processing and interpolating data within this framework. These methods have been created with a focus on practicality. The culmination of this work is showcased in the final chapter, where we venture into the realm of machine learning. Here, we explore the potential applications and implications of lattice ten- sors in machine learning research, underscoring their utility and effectiveness. This exploration not only demonstrates the practical applicability of our proposed methods but also opens new av- enues for research in machine learning, offering fresh perspectives and tools for tackling complex computational problems. In essence, this thesis presents a comprehensive study that bridges the gap between theoretical concepts and practical applications in the approximation of natural phenomena on non-Cartesian grids. Through the introduction of the lattice tensor and associated methodologies, this work con- tributes significantly to the field, providing a robust foundation for future research and development in both computational science and machine learning.
- ItemOpen AccessDanceShala - A Visual Feedback Interface for Dance Learning(2022-08-31) Mukherjee, Suvojit; Alim, Usman; Kenny, Sarah; Katz, Larry; Shekhar Nittala, AdityaDance is a beautiful art form that can be enjoyed by people irrespective of age. One can learn dance from a dance teacher in a dance studio. The visual feedback received in an in-person class from an instructor is one of the best ways to improve dance learning. Sometimes it is not possible for a person to attend dance classes due to time and location constraints. The alternate option for people to learn dance is to attend dance classes online or self-learning with the help of dance games (online video games where a player attempts to follow a pattern of dance steps shown on screen in time to music). Organized remote visual feedback can assist a learner to learn dance in such scenarios. However, online dance classes or dance games may not be sufficient for a new learner to learn dance because the feedback received is not always adequate. To make online dance learning more comprehensive for new learners, a visual feedback interface named ‘DanceShala’ is created which will deliver comparative visual feedback to students after comparing teacher and student movements. In this study, dance movements of the teacher and the student are recorded. After processing the recorded movement data, feedback is generated on the correctness of the student’s movements as compared to the teacher. The visual feedback is displayed through an interface which assists a student to identify the errors made when compared to the teacher video. In the last stage of the study, a survey is administered to understand the user perception about this interface. This research is an interdisciplinary study combining Computer Science, Kinesiology and Dance.
- ItemOpen AccessDecal-maps: Real-time Layering of Decals on Surfaces for Multivariate Visualization(IEEE, 2017-01) Rocha, Allan; Alim, Usman; Silva, Julio Daniel; Sousa, Mario CostaWe introduce the use of decals for multivariate visualization design. Decals are visual representations that are used for communication; for example, a pattern, a text, a glyph, or a symbol, transferred from a 2D-image to a surface upon contact. By creating what we define as decal-maps, we can design a set of images or patterns that represent one or more data attributes. We place decals on the surface considering the data pertaining to the locations we choose. We propose a (texture mapping) local parametrization that allows placing decals on arbitrary surfaces interactively, even when dealing with a high number of decals. Moreover, we extend the concept of layering to allow the co-visualization of an increased number of attributes on arbitrary surfaces. By combining decal-maps,color-maps and a layered visualization, we aim to facilitate and encourage the creative process of designing multivariate visualizations. Finally, we demonstrate the general applicability of our technique by providing examples of its use in a variety of contexts.
- ItemOpen AccessDesigning and Evaluating a Lightweight Video Player for Language Learning(2017) Hu, Sathaporn; Willett, Wesley; Alim, Usman; Chilana, ParmitWatching foreign language videos is a popular and convenient strategy used by many people for learning a new language. However, traditional video players, such as the YouTube player, are not designed to support language learning. We created two video players to explore and to address the issues of using traditional players as a language learning tool. Our players specifically target casual language learners. After evaluating the first player, we found that a traditional player makes it difficult for learners to (1) adjust the level of difficulty, (2) recover missed information, and (3) assess learning progress. We then created the second player to address these issues. The results of the evaluation of the second player demonstrate that people found the player to be helpful for language learning. We also found common usage patterns in the results and opportunities for future improvement.
- ItemOpen AccessEmerging Behavioral and Multi-Modal Biometric Approaches(2018-03-21) Rahman, Md Wasiur; Gavrilova, Marina L.; Alim, Usman; Hagen, Gregory; Alim, Usman R.Emerging behavioral biometrics play a vital role in situations where traditional biometrics may fail to identify the person correctly. This thesis focuses on designing new methods for biometric identification systems using emerging behavioral biometric characteristics such as EEG brain waves and Kinect gait. This research also develops multimodal biometric methods using the combination of emerging behavioral biometric characteristics with physiological biometric characteristics (such as face). The sub-goal of this research is to establish a relationship between a person's mental activity and a system identification accuracy. To fulfill this goal, this research studies overt and covert EEG signals. The second emerging behavioral biometric system is Kinect skeletal gait, where novelty of the work is in considering all possible joint-distance combinations. Finally, two multimodal systems are proposed and developed using the score-level fusion. The performance of the proposed identification systems is evaluated using five publicly available databases.
- ItemOpen AccessEnriched Story Experiences with a New Video Interaction Model(2023-04-28) Ma, Lynshao Celina; Wang, Mea; Alim, Usman; Boyd, Jeffrey EdwinInteractive film is a video storytelling medium lying between films and video games. It traditionally gives users choices at fixed points. This supports non-linear storylines, as the user can influence what happens in the plot by making decisions. However, such films are tightly scripted, and limited in handling users with different needs for interactions. Passive film viewers may find frequent choices disruptive, while active users may feel such choices are shallow and confining. These issues contribute to interactive films being a niche form of entertainment. To lift these restrictions of the standard choice-based interactive film, this thesis proposes a novel anytime interaction film model where users can choose when, and how much to interact. At anytime, users can activate a switching mechanism to enter an interactive mode. This opens a variety of voluntary, plot-relevant interactions in an explorable virtual environment. These broadened interactive possibilities heighten the enjoyment and depth of the system. Furthermore, the complete user experience is personalized by implicitly guiding users to suitable interactions and branching storylines. Interaction guidance streamlines the recommendation of content to adapt to different user preferences. This module is achieved with the online learning of contextual bandits, using domain-specific features. Conversely, a story guidance module performs implicit emotion recognition on the user, as they engage with conversational agents representing story characters. Powered by natural language generation, these agents can discuss any topic the user wishes. Aside from granting an ongoing source of interaction, these agent conversations help model the user’s character preferences. Based on these feelings, suitable storylines for the user can be selected with little authorial scripting. Through user studies and simulations, the worth of these features in enhancing user engagement and autonomy was realized. The new features of environmental exploration and character agents improved user satisfaction. User feedback also suggested this model’s strengthened flexibility was appealing. Furthermore, simulation of the personalization schemes indicated their potential in accommodating different types of users and scenarios. This unique anytime interaction film model eases authorial burden while granting an adaptive user experience, driving a new form of video storytelling.
- ItemOpen AccessevoExplore: Multiscale Visualization for Evolutionary Design Histories(2018-01-26) Kelly, Justin; Jacob, Christian; Alim, Usman; Jacob, Christian; Denzinger, JörgevoExplore is a system designed to assist evolutionary design projects. Built in the Unity 3D game engine and designed with future development and expansion in mind, evoExplore allows the user to collect and visualize data from evolutionary design experiments in 3D. The user is provided the tools needed to breed their own evolving populations, record the results from such evolutionary experiments and then visualize the recorded data as a series of 3D columns and rings representing the recorded evolutionary experiments and their populations over time. evoExplore allows the user to dynamically explore their own evolutionary experiments, as well as those produced by other users, in order to better understand the data produced by these evolutionary systems. In this document we describe the features of evoExplore, the engine it was built in, the use of virtual reality in evoExplore and the contributions our system brings to the field
- ItemOpen AccessExploring Convolutional Neural Networks and Transfer Learning for Oil Sands Drill Core Image Analysis(2021-08-24) Anzum, Fahim; Costa Sousa, Mario; Alim, Usman; John Jacobson Jr., Michael; Osvaldo Trad, Daniel; Zhao, RichardAn accurate permeability estimate is crucial for effectively characterizing the McMurray oil sands for in situ recovery. Such an estimate is critical to inform the best locations for placing wells and pads and accurately forecast future oil production rates. This fact is becoming significantly important as in situ development moves to areas of increasingly complex geology. The traditional methods of estimating permeability largely do not work well in oil sands because of the core disturbance or the fact that the core is filled with immobile bitumen. Moreover, it is expensive to get physical samples from many different depths at many wells, and the experiments carried out in the labs to measure permeability sometimes are not representative. However, permeability can be estimated from different parameters such as mean grain size (MGS), median grain size, and particle size distribution (PSD). This thesis investigates how convolutional neural networks (CNNs) and transfer learning perform when estimating MGS from the oil sands drill core photos. Three preliminary approaches are explored for classifying core photos based on the facies, including (1) the application of transfer learning on the pre-trained VGG-16 CNN model, (2) fine-tuning a few top layers of VGG-16, and (3) the combination of VGG-16 and traditional machine learning (ML) algorithms. Experimental results achieved by these classification models reveal opportunities to extend these approaches for predicting MGS from core photos. Therefore, the three approaches are then investigated using a library of core photographs with known MGS calculated from PSD to see which one works best. Experimental results exhibit good performance in estimating MGS from core photos using the explored approaches. Overall, the investigation supports that the application of CNNs, and transfer learning is feasible in different oil sands drill core image analysis workflows and more advanced research outcomes can be achieved by further exploration of these techniques in the oil sands research domain.
- ItemOpen AccessExploring Immersive Virtual Environments for Well Placement Optimization in Reservoir Models(2017) Cabral Ramos Mota, Roberta; Costa Sousa, Mário; Sharlin, Ehud; Alim, Usman; Parlac, VeraImmersive virtual environments have been considered promising mediums to attend specific demands from the oil and gas industry. In this thesis, we explore immersive technologies’ benefits for the execution of tasks associated with well placement optimization. We present a) an analytical method to perform static connectivity analysis as a proxy for flow simulation, b) an application to support well optimization using our method, and c) an exploration of our application in three immersive environments – a CAVE with a tracked gamepad; a HMD with a tracked gamepad; and a HMD with a leap motion controller – in the search for visualization and interaction techniques that facilitate well placement studies. Based on primary study conducted with reservoir engineers, we provide an examination of the usefulness of our application. We also discuss our findings considering engineers’ preferences as well as the suitability of the different immersive environments for designing and assessing well placement scenarios.
- ItemOpen AccessExploring Temporal Loss Tolerance of Video Codecs for QoE Enhancement in Adaptive Streaming(2021-08-25) Liu, Yang; Wang, Mea; Alim, Usman; Krishnamurthy, DiwakarVideo streaming is not only for entertainment now, but also for lessons, communications, meetings, and even diagnosis. Quality of Experience (QoE) is the most important concerned aspect by content providers and platform providers. Among all the experience, enduring stalls is the most frustrating one. In this thesis, we present Bandwidth-Efficient Temporal Adaptation (BETA) and Temporal Adaptive Streaming over QUIC (TASQ), whose target is to improve the smoothness of video playbacks even under extreme network conditions. We investigate temporal loss tolerance for HEVC and AV1, and apply the optimization on video streaming. Results show a drastic improvement on the smoothness and the QoE.