Browsing by Author "Alim, Usman R."

Now showing 1 - 16 of 16
  • Thumbnail Image
    ItemOpen Access
    A Fast Fourier Transform with Rectangular Output on the BCC and FCC Lattices
    (2009-05) Alim, Usman R.; Möller, Torsten
    This paper discusses the efficient, non-redundant evaluation of a Discrete Fourier Transform on the three dimensional Body-Centered and Face-Centered Cubic lattices. The key idea is to use an axis aligned window to truncate and periodize the sampled function which leads to separable transforms. We exploit the geometry of these lattices and show that by choosing a suitable non-redundant rectangular region in the frequency domain, the transforms can be efficiently evaluated using the Fast Fourier Transform.
  • Thumbnail Image
    ItemOpen Access
    Automated Performance Assessment of Virtual Temporal Bone Dissection
    (2020-07-21) Sachan, Surbhi; Chan, Sonny; Alim, Usman R.; Boyd, Jeffrey Edwin; Forkert, Nils Daniel
    Mastoidectomy is a surgical procedure in which a portion of the temporal bone is removed by using fine microsurgical skills. Development of virtual reality simulators with high-fidelity visual, auditory, and force feedback has allowed trainees to learn this skill in a safe environment without the limitations associated with the traditional way of learning, i.e., cadaveric specimens. However, without an automatic feedback mechanism, an expert's presence is required to assess the performance, placing a heavy burden on their time. This investigation focuses on automating the performance evaluation obviating the need for an expert's time. This is accomplished by automating the criteria based on a well-established and validated assessment instrument known as the Welling Scale, to score the mastoidectomy performed on a virtual surgery simulator. Image processing algorithms are devised and run on the output of the virtual surgery to automatically score these criteria. The criteria are described in terms of four functional categories: Identification, Skeletonization, Intactness and No cells. Algorithms are devised for each of these categories. This work further validates the accuracy of these algorithms by doing a study where these criteria are evaluated by two experts, as well as the work done in this thesis. The results of the study show that automatic performance assessment of virtual mastoidectomy surgery is feasible.
  • Thumbnail Image
    ItemOpen Access
    Compactly Supported Biorthogonal Wavelet Bases on the Body Centered Cubic Lattice
    (Wiley, 2017-06) Horacsek, Joshua J.; Alim, Usman R.
    In this work, we present a family of compact, biorthogonal wavelet filter banks that are applicable to the 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.
  • Thumbnail Image
    ItemOpen Access
    Compressive Volume Rendering
    (Wiley, 2015-05) Liu, Xiaoyang; Alim, Usman R.
    Compressive rendering refers to the process of reconstructing a full image from a small subset of the rendered pixels, thereby expediting the rendering task. In this paper, we empirically investigate three image order techniques 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. The latter techniques exploit smoothness properties of the rendered image; the second technique recovers the missing pixels via a total variation minimization procedure while the third technique incorporates a smoothness prior in a variational reconstruction framework employing interpolating cubic B-splines. We compare and contrast the three 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 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.
  • Thumbnail Image
    ItemOpen Access
    Distributed Video Transcoding for Live and Interactive Multimedia Applications
    (2019-08-26) Sameti, Sajad; Krishnamurthy, Diwakar; Wang, Mea; Hemmati, Hadi; Alim, Usman R.
    One of the major challenges of video service providers is offering high Quality of Experience (QoE). Satisfying high QoE is a challenging problem given the diversity in end-device characteristics and network conditions as well as the increased popularity of live and interactive streaming applications. For a user to have acceptable QoE, the streaming services need to adapt video quality to device specifications and network quality at runtime. Runtime adaptation, i.e., transcoding, is a challenging problem due to the computationally intensive nature of the transcoding process. In particular, novel techniques are needed for maintaining the real-time property of a video stream in spite of this high computational demand. This thesis proposes Stride and Contrast, two different distributed transcoding approaches that address this challenge. We outline the design and performance optimization of these approaches. Results show that our approaches outperform other similar techniques used in practice and proposed by other researchers.
  • Thumbnail Image
    ItemOpen Access
    Emerging 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.
  • Thumbnail Image
    ItemOpen Access
    Illustrative Multivariate Visualization for Geological Modelling
    (Wiley, 2018-07-10) Rocha, Allan; Mota, Roberta Cabral Ramos; Hamdi, Hamidreza; Alim, Usman R.; Sousa, Mario Costa
    In this paper, we present a novel illustrative multivariate visualization for geological modelling to assist geologists and reservoir engineers in visualizing multivariate datasets in superimposed representations, in contrast to the single-attribute visualizations supported by commercial software. Our approach extends the use of decals from a single surface to 3D irregular grids, using the layering concept to represent multiple attributes. We also build upon prior work to augment the design and implementation of different geological attributes (namely, rock type, porosity, and permeability). More specifically, we propose a new sampling strategy to generate decals for porosity on the geological grid, a hybrid visualization for permeability which combines 2D decals and 3D ellipsoid glyphs, and a perceptually-based design that allows us to visualize additional attributes (e.g., oil saturation) while avoiding visual interference between layers. Furthermore, our visual design draws from traditional geological illustrations, facilitating the understanding and communication between interdisciplinary teams. An evaluation by domain experts highlights the potential of our approach for geological modelling and interpretation in this complex domain.
  • Thumbnail Image
    ItemOpen Access
    Improving Quality of Experience (QoE) of Dynamic Adaptive Streaming (DAS) Systems
    (2019-09-19) James, Cyriac; Wang, Mea; Krishnamurthy, Diwakar; Williamson, Carey L.; Alim, Usman R.; Liu, Yang
    Dynamic Adaptive Streaming (DAS) systems dominate today's video streaming over the Internet, and operate by adapting video quality based on network throughput variation using discrete quality levels. Despite their popularity, it lacks an effective adaptation that minimizes stalls and quality switches while maximizing visual quality, especially when available bandwidth varies. The conventional approach to adaptation is to make a decision on the next video segment quality based on prior throughput measurements. This approach is not robust to bandwidth fluctuation at small time scales, which can consequently lead to stalls, bandwidth waste, and unstable quality, mainly due to the inability to mitigate significant bandwidth reduction during the segment download. MultiPath TCP (MPTCP) is an emerging paradigm that could offer significant benefits to video streaming by harnessing bandwidth from multiple network interfaces, in particular on mobile and desktop devices with support for both WiFi and cellular networks. We first investigate this off-the-shelf solution to improve video streaming performance by harvesting additional bandwidth over always or intermittently available secondary link under different bandwidth variability conditions. Our measurement study yields mixed results. While beneficial to user experience when primary link bandwidth is unstable or constrained, MPTCP may not offer any advantage otherwise, and sometimes could be detrimental. We then propose BETA – Bandwidth-Efficient Temporal Adaptation, an agile approach that allows DAS players to refine the quality level within video segments on the fly, according to the actual bandwidth conditions experienced while downloading each segment. We define a new DAS-oriented transmission order of video frames within segments that facilitates decodability of partial frames, and paves the way for changing the paradigm from discrete to continuous bitrate ladders for DAS. BETA can work with any adaptation algorithm that runs on a DAS player to significantly improve robustness and efficiency in dynamic network environments and for low-latency streams, as well as dramatically reduce content storage and encoding infrastructure requirements.
  • Thumbnail Image
    ItemOpen Access
    Meta-Feature Taxonomy for Supporting Automatic Machine Learning
    (2019-12-23) Davies, Cooper; Maurer, Frank; Denzinger, Jörg; Jacob, Christian; Alim, Usman R.; Oehlberg, Lora A.
    Many automatic machine learning (AutoML) libraries have been developed recently, meeting public demand for more machine learning tools which can be used without an expert. A common tactic illicited by these frameworks is to initially generate meta-features which are then used as an initial heuristic for further evaluation in recent AutoML frameworks. In this thesis we provide a systematic categorization of meta-features in the AutoML literature. Current implementations of automatic machine learning frameworks fail to provide reasoning for meta-feature selection, and a taxonomic categorization is needed. Our approach reviewed current AutoML frameworks and created a taxonomy of five categories into which any meta-feature can be categorized. We have created a general framework with which any currently used meta-features can be described, as well as demonstrate some scenarios for their applications. Additionally, a runtime analysis of the wall-clock time required for meta-feature generation is provided for 18 data collections found in previous CHALearn AutoML competitions, which took between 0:10:26.9, and 98:43:46.5. Additionally we found that a sample percentage of 0.1 is sufficient for use in Sample Variant Landmark Meta-Feature generation when using the Nearest Neighbour, Elite Nearest Neighbour, Best Decision Node, and Random Decision Node Landmarks which indicates potential use as meta-features in AutoML.
  • Thumbnail Image
    ItemOpen Access
    Modelling Natural Phenomenon with Reaction-Diffusion
    (2020-01-22) Ringham, Lee; Prusinkiewicz, Przemysław; Alim, Usman R.; Jacob, Christian J.
    Procedural methods provide an algorithmic way to produce textures for use in computer graphics. One such method, reaction-diffusion, is a powerful mathematical approach that describes natural pattern formation in terms of chemicals known as morphogens. This thesis describes LRDS, an environment for authoring reaction-diffusion models directly on arbitrary surfaces. Morphogens, their behaviours, and the domain in which they reside can be quickly and easily defined. By performing computation on the GPU, the pattern forming simulation can be interacted with in real-time, facilitating productivity and experimentation. Four case studies are presented. The first is a simulation of ladybug pigmentation patterns. The second is a simulation of pigmentation patterns seen on the body of snakes. The third study looks at flower petal pattern modelling. Lastly, a biologically-motivated model of the autoimmune disease psoriasis is presented.
  • Thumbnail Image
    ItemOpen Access
    Multiresolution by Repeated Invertible Averaging - With Applications in Digital Earth
    (2019-11) Alderson, Troy F.; Samavati, Faramarz; Stefanakis, Emmanuel; Akleman, Ergun; Prusinkiewicz, Przemysław W.; Alim, Usman R.
    In this thesis, we present a general-purpose, arbitrary-degree framework for the multiscale representation of various types of graphics objects. These include curves in Euclidean and non-Euclidean spaces (particularly the surfaces of spheres and ellipsoids) as well as polygonal mesh surfaces. The core framework, which operates on curves, is based on simple yet fundamental modifications to the Lane-Riesenfeld algorithm and its generalizations. The algorithm’s averaging step is replaced with invertible variants, defining a repeated invertible averaging approach that supports a class of subdivision and reverse subdivision methods (including those that produce B-Spline curves). These averaging steps and their inverses are defined in terms of sequences of two-point interpolations between neighbouring vertices, which can be easily generalized to several different spaces and manifolds. In addition to developing this core framework, we explore different applications and generalizations of the approach. In particular, we concentrate on applications to Digital Earth, where spherical and ellipsoidal curves can be used to represent geospatial vectors (e.g. nation boundaries, road networks). We use multiscale representations of geospatial vector data to develop fast algorithms for offsetting queries and inside/outside tests. A fast offsetting algorithm for rasterized vectors in a DGGS is also presented. Generalizations of the approach include a modification that allows our framework to produce multiscale NURBS curves on the sphere and ellipsoid. This is accomplished by incorporating vertex weights into the interpolation parameters of individual operations, preserving the framework's generalizability. We additionally present a generalization of our framework to the multiscale representation of polygonal meshes. Similarly to the curve case, our framework for surfaces is defined in terms of local mesh operations that involve only direct vertex neighbours. Smooth reverse and non-uniform surface subdivisions are additionally supported.
  • Thumbnail Image
    ItemOpen Access
    Operator Discretization in Shift-Invariant Spaces
    (2014-06) Alim, Usman R.
  • Thumbnail Image
    ItemOpen Access
    Optically Illusive Architecture (OIA): A Design Paradigm Attuned Towards Viewpoints
    (2020-12) Hosseini, Seyed Vahab; Alim, Usman R.; Taron, Joshua M.; Oehlberg, Lora A.; Johnson, Jason S.
    Human beings have historically recorded events of their surrounding world by means of drawing. Likewise, architects and designers communicate their ideas within a range of representational methods. No single instance of these methods, either in the form of orthographic projections---also known as descriptive geometry---or perspectival representation, can address all questions regarding the design, but as a whole, they demonstrate a comprehensive range of information about the building or object they intend to represent. This explicates an inevitable degree of deficiency in representation, regardless of its type. In addition, perspective-based optical illusions manipulate our spatial perception by deliberately misrepresenting the reality. In this regard, they are not fresh concepts to architectural representation. In this thesis I propose Optically Illusive Architecture (OIA); a viewpoint-sensitive design paradigm whose concept derives from the gap between representational limitations and physical reality. Also, results of this design paradigm deliver specific messages to certain privileged point(s) in the space. OIA casts light to an undeniable role of viewpoints in designing architectural spaces. The idea is to establish a methodology in a way that the deficiency of current representational techniques---manifested as specific thread of optical illusions---flourishes into thoughtful results embodied as actual architectural spaces. Within this design paradigm, I define a framework to be able to effectively analyze its precedents, generate new space and evaluate their efficiencies. Moreover, the framework raises a hierarchical set of questions to differentiate OIA from a visual gimmick. Additional contributions of this thesis are generating two optically illusive architectures, as well as a new method of illusory representation. Furthermore, I analyze the generated OIA environments, by conducting empirical studies using Virtual Reality (VR) technology. These studies demonstrate design performance, and the public’s ability to engage and interact with OIA spaces, prior to the actual fabrication of the structures.
  • Thumbnail Image
    ItemOpen Access
    Simulation of a Kelp Forest using Smoothed Particle Hydrodynamics
    (2021-01-22) Bloor, Cordell G. M.; Prusinkiewicz, Przemysław W.; Natale, Giovanniantonio; Alim, Usman R.
    In the wave-swept waters off the Pacific coasts, the giant kelp, Macrocystis pyrifera, both shapes its environment and is shaped by it. Kelp forests are some of the most important ecosystems on Earth, yet these vibrant, underseas environments are difficult to appreciate from land, or even the ocean’s surface. This work explores methods to model and animate kelp forests in software, bidirectionally coupling a fluid simulation based on Weakly Compressible Smoothed Particle Hydrodynamics (WCSPH) with a model of kelp fronds defined by articulated rigid bodies, with appropriate boundary conditions for the simulation region.
  • Thumbnail Image
    ItemOpen Access
    Sub-band Coding of Hexagonal Images
    (2019-09-16) Rashid, MD Mamunur; Alim, Usman R.; Yanushkevich, Svetlana N.; Wang, Mea
    According to the circle-packing theorem, the packing efficiency of a hexagonal lattice is higher than an equivalent square tessellation. Consequently, in several contexts, hexagonally sampled images compared to their Cartesian counterparts are better at preserving information content. In this thesis, novel mapping techniques alongside the wavelet compression scheme are presented for hexagonal images. Specifically, we introduce two tree-based coding schemes, referred to as SBHex (spirally-mapped branch-coding for hexagonal images) and BBHex (breadth-first block-coding for hexagonal images). Both of these coding schemes respect the geometry of the hexagonal lattice and yield better compression results. Our empirical results show that the proposed algorithms for hexagonal images produce better reconstruction quality at lower bits-per-pixel values compared to the tree-based coding counterparts for the Cartesian grid.
  • Thumbnail Image
    ItemOpen Access
    Toward High Quality Gradient Estimation on Regular Lattices
    (IEEE, 2011-04) Hossain, Zahid; Alim, Usman R.; Möller, Torsten
    In this paper, we present two methods for accurate gradient estimation from scalar field data sampled on regular lattices. The first method is based on the multi-dimensional Taylor series expansion of the convolution sum and allows us to specify design criteria such as compactness and approximation power. The second method is based on a Hilbert space framework and provides a minimum error solution in the form of an orthogonal projection operating between two approximation spaces. Both methods lead to discrete filters which can be combined with continuous reconstruction kernels to yield highly accurate estimators as compared to the current state of the art. We demonstrate the advantages of our methods in the context of volume rendering of data sampled on Cartesian and Body-Centered Cubic lattices. Our results show significant qualitative and quantitative improvements for both synthetic and real data, while incurring a moderate pre-processing and storage overhead.