UMD Theses and Dissertations
Permanent URI for this collectionhttp://hdl.handle.net/1903/3
New submissions to the thesis/dissertation collections are added automatically as they are received from the Graduate School. Currently, the Graduate School deposits all theses and dissertations from a given semester after the official graduation date. This means that there may be up to a 4 month delay in the appearance of a given thesis/dissertation in DRUM.
More information is available at Theses and Dissertations at University of Maryland Libraries.
Browse
5 results
Search Results
Item Immersive Visual Analytics of Wi-Fi Signal Propagation and Network Health(2023) Rowden, Alexander R; Varsnhney, Amitabh; Computer Science; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)e are immersed in waves of information. This information is typically transmitted as radio waves in many protocols and frequencies, such as WiFi, Bluetooth, and Near-Field Communications (NFC). It carries vital information such as health data, private messages, and financial records. There is a critical need for systematic and comprehensive visualization techniques to facilitate seamless, resilient, and secure transmission of these signals. Traditional visualization techniques are not enough because of the scale of these datasets. In this dissertation, we present three novel contributions that leverage advances in volume rendering and virtual reality (VR): (a) an outdoor volume-rendering visualization system that facilitates large-scale visualization of radio waves over a college campus through real-time programmable customization for analysis purposes, (b) an indoor, building-scale visualization system that enables data to be collected and analyzed without occluding the user's view of the environment, and (c) a systematic user study with 32 participants which shows that users perform analysis tasks well with our novel visualizations. In our outdoor system, we present the Programmable Transfer Function. Programmable Transfer Functions offer the user a way to replace the traditional transfer function paradigm with a more flexible and less memory-demanding alternative. Our work on indoor WiFi visualization is called WaveRider. WaveRider is our contribution to indoor-modeled WiFi visualization using a virtual environment. WaveRider was designed with the help of expert signal engineers we interviewed to determine the needs of the visualization and who we used to evaluate the application. These works provide a solid starting point for signal visualization as our networks transition to more complex models. Indoor and outdoor visualizations are not the only dichotomy in the realm of signal visualization. We are also interested in visualizations of modeled data compared to visualization of data samples. We have also explored designs for multiple sample-based visualizations and conducted a formal evaluation where we compare these to our previous model-based approach. This analysis has shown that visualizing the data without modeling improves user confidence in their analyses. In the future, we hope to explore how these sample-based methods allow more routers to be visualized at once.Item Learning and Composing Primitives for the Visual World(2023) Gupta, Kamal; Shrivastava, Abhinav; Davis, Larry; Computer Science; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Compositionality is at the core of how humans understand and create visual data. In order for the computational approaches to assist humans in creative tasks, it is crucial for them to understand and perform composition. The recent advances in deep generative models have enabled us to convert noise to highly realistic scenes. However, in order to harness these models for building real-world applications, I argue that we need to be able to represent and control the generation process with the composition of interpretable primitives. In the first half of this talk, I’ll discuss how deep models can discover such primitives from visual data. By playing a cooperative referential game between two neural network agents, we can represent images with discrete meaningful concepts without supervision. I further extend this work for applications in image and video editing by learning a dense correspondence of primitives across images. In the second half, I’ll focus on learning how to compose primitives for both 2D and 3D visual data. By expressing the scenes as an assembly of smaller parts, we can easily perform generation from scratch or from partial scenes as input. I’ll conclude the talk with a discussion of possible future directions and applications of generative models, and how we can better enable users to guide the creative process.Item Dense 3D Reconstructions from Sparse Visual Data(2022) Hu, Tao; Zwicker, Matthias; Computer Science; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)3D reconstruction, the problem of estimating the complete geometry or appearance of objects from partial observations (e.g., several RGB images, partial shapes, videos), serves as a building block in many vision, graphics, and robotics applications such as 3D scanning, autonomous driving, 3D modeling, augmented reality (AR) and virtual reality (VR). However, it is very challenging for machines to recover 3D geometry from such sparse data due to occlusions, and irregularity and complexity of 3D objects. To solve these, in this dissertation, we explore learning-based 3D reconstruction methods for different 3D object representations on different tasks: 3D reconstructions of static objects and dynamic human body from limited data. For the 3D reconstructions of static objects, we propose a multi-view representation of 3D shapes, which utilizes a set of multi-view RGB images or depth maps to represent a 3D shape. We first explore the multi-view representation for shape completion tasks and develop deep learning methods to generate dense and high-resolution point clouds from partial observations. Yet one problem with the multi-view representation is the inconsistency among different views. To solve this problem, we propose a multi-view consistency optimization strategy to encourage consistency for shape completion in inference stage. Third, the extension of multi-view representation for dense 3D geometry and texture reconstructions from single RGB images will be presented. Capturing and rendering realistic human appearances under varying poses and viewpoints is an important goal in computer vision and graphics. In the second part, we will introduce some techniques to create 3D virtual human avatars with limited data (e.g., videos). We propose implicit representations of motion, texture, and geometry for human modeling, and utilize neural rendering techniques for free view synthesis of dynamic articulated human body. Our learned human avatars are photorealistic and fully controllable (pose, shape, viewpoints, etc.), which can be used in free-viewpoint video generation, animation, shape editing, telepresence, and AR/VR. Our proposed methods can learn end-to-end 3D reconstructions from 2D image or video signals. We hope these learning-based methods will assist in perceiving and reconstructing the 3D world for future AI systems.Item Computer Graphics Based Optical Tracking for Hypersonic Free-Flight Experiments(2019) Starshak, William; Laurence, Stuart; Aerospace Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Aerodynamic measurement in hypersonic short-duration facilities – facilities with test times shorter than 10 milliseconds – is a topic of ongoing research. Standard force-balance approaches cannot handle the short test-time or the flow-initiating shock wave. Experimentalists have developed alternative techniques; but these techniques deliver merely adequate results at the cost of significant operational and – especially – calibration complexity. Recently, Laurence, et al. proposed using high-speed shadowgraph imaging and edge fitting (matching the visualized edge to an analytic equation for that edge) to make high-precision free-flight measurements of capsules. This new technique promised equivalent accuracy to existing techniques with far less pre-test calibration. The technique as developed, however, was limited to simple shapes in 2D motion. This thesis presents a generalization of the edge-fitting concept. Using the correspondence between a model's orientation and its silhouette, the trajectory of a model may be tracked to $1 \, \rm \mu m$ positional and $0.01^{\circ}$ angular accuracy. The silhouette is generated using computer-graphics techniques based upon a 3D mesh of the model's surface geometry. Consequently, the proposed technique is general to the model shape, the number of models, the properties of the camera imaging the experiment, and the number of cameras. Using the technique, we measured the hypersonic aerodynamics of a sphere, a blunt sphere-cone capsule, a lifting-body spacecraft, and the University of Maryland Testudo. In addition, multiple-camera and multiple-body tracking capability is demonstrated with an experiment investigating the dynamics of a breaking-up satellite. Results show that the method achieves accuracy comparable to or better than existing techniques with a simpler experimental procedure.Item Single-View 3D Reconstruction of Animals(2017) Kim, Angjoo; Jacobs, David W; Computer Science; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Humans have a remarkable ability to infer the 3D shape of objects from just a single image. Even for complex and non-rigid objects like people and animals, from just a single picture we can say much about its 3D shape, configuration and even the viewpoint that the photo was taken from. Today, the same cannot be said for computers – the existing solutions are limited, particularly for highly articulated and deformable objects. Hence, the purpose of this thesis is to develop methods for single-view 3D reconstruction of non-rigid objects, specifically for people and animals. Our goal is to recover a full 3D surface model of these objects from a single unconstrained image. The ability to do so, even with some user interaction, will have a profound impact in AR/VR and the entertainment industry. Immediate applications are virtual avatars and pets, virtual clothes fitting, immersive games, as well as applications in biology, neuroscience, ecology, and farming. However, this is a challenging problem because these objects can appear in many different forms. This thesis begins by providing the first fully automatic solution for recovering a 3D mesh of a human body from a single image. Our solution follows the classical paradigm of bottom-up estimation followed by top-down verification. The key is to solve for the mostly likely 3D model that explains the image observations by using powerful priors. The rest of the thesis explores how to extend a similar approach for other animals. Doing so reveals novel challenges whose common thread is the lack of specialized data. For solving the bottom-up estimation problem well, current methods rely on the availability of human supervision in the form of 2D part annotations. However, these annotations do not exist in the same scale for animals. We deal with this problem by means of data synthesis for the case of fine-grained categories such as bird species. There is also little work that systematically addresses the 3D scanning of animals, which almost all prior works require for learning a deformable 3D model. We propose a solution to learn a 3D deformable model from a set of annotated 2D images with a template 3D mesh and from a few set of 3D toy figurine scans. We show results on birds, house cats, horses, cows, dogs, big cats, and even hippos. This thesis makes steps towards a fully automatic system for single-view 3D reconstruction of animals. We hope this work inspires more future research in this direction.