The Totally Digital Imaging System (TDIS) is based on system requirements information from the Radiology Department, University of Arizona Health Science Center. This dissertation presents the design of this complex system, the TDIS specification, the system performance requirements, and the evaluation of the system using the computer simulation programs. Discrete event simulation models were developed for the TDIS subsystems, including an image network, imaging equipment, storage migration algorithm, data base archive system, and a control and management network. The simulation system uses empirical data generation and retrieval rates measured at the University Medical Center hospital. The entire TDIS system was simulated in Simscript II.5 using a VAX 8600 computer system. Simulation results show the fiber optical image network to be suitable, however, the optical disk storage system represents a performance bottleneck.

This dissertation presents image processing methods designed to enhance images obtained by angiography, and applied image analysis methods to quantify the vascular diameter. An iterative, non-linear enhancement technique is described for enhancing the edges of blood vessels in unsubtracted angiographic images. The technique uses a median filter and the point spread function of the imaging system to increase the resolution of the image while keeping down noise. Evaluation of the images by radiologists showed that they preferred the processed images over the unprocessed images. Also described is a heuristic, recursive, vessel tracking algorithm. The tracker is intended for use with digital subtraction angiography images. The vascular system is characterized by a tree data structure. Tree structures are inherently recursive structures and thus recursive programming languages are ideally suited for building and describing them. The tracker uses a window to follow the centerlines of the vessels and stores parameters describing the vessels in nodes of a binary tree. Branching of the vascular tree is handled automatically. A least squares fit of a cylindrical model to intensity profiles of the vessel is used to estimate vessel diameter and other parameters. The tracker is able to successfully track vessels with signal-to-noise ratios down to about 4. Several criteria are applied to distinguish between vessel and noise. The relative accuracy of the diameter estimate is about 3% to 8% for a signal-to-noise ratio of 10; the absolute accuracy depends on the magnification (mm per sample). For the clinically significant case of a 25% stenosis (narrowing of the vessel), the absolute error in estimating the percent stenosis is 3.7% of the normal diameter and the relative error is 14.8%. This relative error of 14.8% is a substantial improvement over relative errors of 30% to 70% produced by other methods.

University of Macau / Faculty of Science and Technology. / Department of Computer and Information Science

The retrieval of digital images is hindered by the semantic gap. The semantic gap is the disparity between a user's high-level interpretation of an image and the information that can be extracted from an image's physical properties. Content based image retrieval systems are particularly vulnerable to the semantic gap due to their reliance on low-level visual features for describing image content. The semantic gap can be narrowed by including high-level, user-generated information. High-level descriptions of images are more capable of capturing the semantic meaning of image content, but it is not always practical to collect this information. Thus, both content-based and human-generated information is considered in this work. A content-based method of retrieving images using a computational model of visual attention was proposed, implemented, and evaluated. This work is based on a study of contemporary research in the field of vision science, particularly computational models of bottom-up visual attention. The use of computational models of visual attention to detect salient by design regions of interest in images is investigated. The method is then refined to detect objects of interest in broad image databases that are not necessarily salient by design. An interface for image retrieval, organization, and annotation that is compatible with the attention-based retrieval method has also been implemented. It incorporates the ability to simultaneously execute querying by image content, keyword, and collaborative filtering. The user is central to the design and evaluation of the system. A game was developed to evaluate the entire system, which includes the user, the user interface, and retrieval methods. / by Liam M. Mayron. / Thesis (Ph.D.)--Florida Atlantic University, 2008. / Includes bibliography. / Electronic reproduction. Boca Raton, FL : 2008 Mode of access: World Wide Web.

Computer simulation of human foot models can be widely used in areas such as surgery simulation, footwear design, gait analysis, etc. Nevertheless, due to the complexity in the modeling of human foot, not much work in the modeling and simulating of human foot model has been reported. In this thesis, we propose an anatomy-based approach for modeling and animating human foot. The anatomical model is divided into layers including skin, muscle, tendon and skeleton. In order to generate animation of the foot model, foot bone structure is transformed according to the foot motions. This induces the deformation of the muscle, tendon and consequently the skin layer. Physics-based approaches are adopted to deform the muscles and tendons. / In this thesis, we adopt the boundary element method (BEM) with linear boundary elements to model deformation of the muscles. Under our investigation, computing the deformation of models with linear boundary elements of BEM is significantly faster than the BEM with constant boundary elements. It also performs better than the finite element method (FEM) under most circumstances. The role of the tendons on determining deformation of the skin layer is also presented. To allow fast deformation, the axial deformation technique is adopted. Using the axial deformation technique, the shape of the axial curve is adjusted to control the deformation of tendon. Two approaches for updating the axial curve are presented. One approach deforms the axial curve of the tendon using geometric technique based on information from image data. Another approach updates the axial curve based on physical properties of the tendon using the mass-spring system. Each vertex of the skin layer is associated with the underlying muscles, tendons and skeleton. When the underlying muscle, tendon and skeleton layers change their shapes, the skin layer is deformed accordingly. Experimental results illustrated that the visual realism of a foot model is enhanced by considering the changes in foot tendons in the deformation of skin layer. / Tang, Yuk Ming. / "January 2007." / Adviser: Kin-Chuen Hui. / Source: Dissertation Abstracts International, Volume: 68-08, Section: B, page: 5372. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2007. / Includes bibliographical references (p. 184-194). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract in English and Chinese. / School code: 1307.

Previous perspective reparameterizations approximate the aliasing errors for the points in a 2D space---the points along the view direction (i.e. horizontal plane in view space), from a geometric intuition. There's no explicit way to qualitatively analyze the aliasing elsewhere. We thus derive the global representation of aliasing functions based on the solid mathematical analysis. Using the global representation, we propose the Direction-adaptive Perspective Shadow Maps (DirPSMs) to offer enhanced shadow rendering in real-time applications. / Shadows are essential for the realism of computer-generated images, which dramatically enhance our perception of virtual scenes by providing useful visual hints. Even tremendous advances in graphics rendering algorithms and programmable GPUs have made real-time and photo-realistic rendering a reality, the synthesis of realistic shadowing effects is still challenging and computationally intensive. Shadow mapping is one of the most popular algorithms for real-time shadow rendering, which has been extensively adopted in real-time applications by its generality and efficiency. However, shadow mapping suffers from the inherent aliasing problems such as jagged shadow boundaries and incorrect self-shadowing due to the image-based nature. In this thesis, we present several shadow rendering techniques to render anti-aliased shadows via shadow mapping especially for large-scale and dynamic scenes. / The Parallel-Split Shadow Maps (PSSMs) scheme is first proposed for realistic and real-time shadow rendering towards next generation of three-dimensional computer games. This scheme splits the view frustum into multiple parts using clip planes parallel to the view plane, and then generates multiple smaller shadow maps for the split parts. A fast and robust split strategy based on the analysis of shadow map aliasing is developed, to produce a moderate aliasing distribution over the whole depth range. Hardware-specific accelerations are also implemented and tested to remove the extra rendering passes caused by using multiple shadow maps. / These techniques are intuitive to implement without using complex data structures. They are designed to improve both quality and performance of shadow rendering in large-scale and dynamic scenes. / With the observation that the sampling density at the light in the post-perspective space better accommodates the requirements for the reconstruction of shadowed images, perspective reparameterization techniques warp shadow maps using the perspective transform to reduce aliasing errors for the objects near the viewer. One limitation in this line of research is that the approximated representation of aliasing errors is only valid for the ideal case in which the light and view directions are orthogonal. We thus establish the generalized perspective reparameterization functions to adaptively adjust the warping strength as the light and/or viewer moves. With the direction-dependent representations of aliasing errors, we subsequently develop three enhanced perspective reparameterizations to better handle general cases, including Generalized Linear Perspective Reparameterization (GLPR), Generalized Minimum-Norm Perspective Reparameterization (GMNPR) and Focus-preserving Trapezoidal Shadow Maps (FTSMs). / Zhang, Fan. / "August 2007." / Adviser: Hanqiu Sun. / Source: Dissertation Abstracts International, Volume: 69-02, Section: B, page: 1126. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2007. / Includes bibliographical references (p. 147-155). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract in English and Chinese. / School code: 1307.

Acquiring 3-D motion of a camera from image sequences is one of the key components in a wide range of applications such as human computer interaction. Given the 3-D structure, the problem of camera motion recovery can he solved using the model-based approaches, which are well-known and have good performance under a controlled environment. If prior information on the scene is not available, traditional Structure from Motion (SFM) algorithms, which simultaneously estimate the scene structure and pose information, are required. The research presented in this thesis belongs to a different category: Motion from Motion (MFM), in which the main concern is the camera position and orientation. To be more precise, MFM algorithms have the capability of estimating 3-D camera motion directly from 2-D image motion without the explicit reconstruction of the scene structure, even though the 3-D model structure is not known in prior. As keeping track of the structural information is no longer required, putting these types of algorithms into real applications is relatively easy and convenient. / It is demonstrated in the experiments that the proposed algorithms are efficient, stable and accurate compared to several existing approaches. Furthermore, they have been put into applications such as mixed reality, virtual reality, robotics and super-resolution to show their performance in real situations. / The objective of this thesis is to develop a high-speed recursive approach that tackles the MFM problem. On the way to the final goal, a series of methods, each having its own strengths and characteristics, have been studied. (1) The first algorithm computes the camera pose from a monocular image sequence. The trifocal tensor is incorporated into the Extended Kalman Filter (EKF) formulation. The step of computing the 3-D models can thus be eliminated. (2) The proposed approach is then extended to the recovery of motion from a stereo image sequence. By applying the trifocal tensor to a stereo vision framework, the trifocal constraint becomes more robust and is not likely to be degenerate. In addition, the twist motion model is adopted to parameterize the 3-D motion. It does not suffer from singularities as Euler angles, and is minimal as opposed to quaternion and the direct use of rotation matrix. (3) The third method introduces the Interacting Multiple Model Probabilistic Data Association Filter (IMMPDAF) to the MFM problem. The Interacting Multiple Model (IMM) technique allows the existence of more than one dynamic system and in return leads to improved accuracy and stability even under abrupt motion changes. The Probabilistic Data Association (PDA) framework makes the automatic selection of measurement sets possible, resulting in enhanced robustness to occlusions and moving objects. As the PDA associates stereo correspondences probabilistically, the explicit establishment of stereo matches is not necessary except during initialization, and the point features present in the outer region of the stereo images can be utilized. / Yu, Ying Kin. / "July 2007." / Adviser: Wong Kin Hong. / Source: Dissertation Abstracts International, Volume: 69-02, Section: B, page: 1125. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2007. / Includes bibliographical references (p. 120-130). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract in English and Chinese. / School code: 1307.

But haze removal is highly challenging due to its mathematical ambiguity, typically when the input is merely a single image. In this thesis, we propose a simple but effective image prior, called dark channel prior, to remove haze from a single image. The dark channel prior is a statistical property of outdoor haze-free images: most patches in these images should contain pixels which are dark in at least one color channel. Using this prior with a haze imaging model, we can easily recover high quality haze-free images. Experiments demonstrate that this simple prior is powerful in various situations and outperforms many previous approaches. / Haze is a natural phenomenon that obscures scenes, reduces visibility, and changes colors. It is an annoying problem for photographers since it degrades image quality. It is also a threat to the reliability of many applications, like outdoor surveillance, object detection, and aerial imaging. So removing haze from images is important in computer vision/graphics. / Speed is an important issue in practice. Like many computer vision problems, the time-consuming step in haze removal is to combine pixel-wise constraints with spatial continuities. In this thesis, we propose two novel techniques to solve this problem efficiently. The first one is an unconventional large-kernel-based linear solver. The second one is a generic edge-aware filter which enables real-time performance. This filter is superior in various applications including haze removal, in terms of speed and quality. / The human visual system is able to perceive haze, but the underlying mechanism remains unknown. In this thesis, we present new illusions showing that the human visual system is possibly adopting a mechanism similar to the dark channel prior. Our discovery casts new insights into human vision research in psychology and physiology. It also reinforces the validity of the dark channel prior as a computer vision algorithm, because a good way for artificial intelligence is to mimic human brains. / He, Kaiming. / Adviser: Xiaoou Tang. / Source: Dissertation Abstracts International, Volume: 73-06, Section: B, page: . / Thesis (Ph.D.)--Chinese University of Hong Kong, 2011. / Includes bibliographical references (leaves 131-138). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [201-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract also in Chinese.

Five algorithms in different applications are proposed in this thesis. All of them are formulated as smoothness based labeling problems, including single image segmentation, video object cutout, image/video completion, image denoising, and image matting. According to different definitions, different optimization approaches are used in these algorithms. In single image segmentation and video object cutout, the graph-cut algorithms are used; in image/video completion, belief propagation is used; and in image denoising and image matting, closed form optimization is implemented. / Many applications in computer vision can be formulated as labeling problems of assigning each pixel a label where the labels represent some local quantities. If all pixels are regarded as independent, i.e., the label of each pixel has nothing to do with the labels of other pixels, such labeling problems are seriously sensitive to noise. On the other hand, for applications in videos, if the inter-frame information is neglected, the performance of the algorithms will be degraded. / Successful performance of the five proposed algorithms, with comparisons to related methods, demonstrates that the proposed models of the labeling problems using the smoothness-based priors work very well in these computer vision applications. / Such labeling problems with smoothness-based priors can be solved by minimizing a Markov energy. According to different definitions of the energy functions, different optimization tools can be used to obtain the results. In this thesis, three optimization approaches are used due to their good performance: graph cuts, belief propagation, and optimization with a closed form solution. / To improve results of these labeling problems, smoothness-based priors can be enforced in the formulations. For a single image, the smoothness is the spatial coherence, which means that spatially close pixels trend to have similar labels. For a video, an additional temporal coherence is enforced, which means that the corresponding pixels in different frames should have similar labels. The spatial coherence constraint makes algorithms robust to noise and the temporal coherence constraint utilizes the inter-frame information for better video-based applications. / Chen, Shifeng. / Adviser: Liu Jian Zhuang. / Source: Dissertation Abstracts International, Volume: 70-06, Section: B, page: 3594. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2008. / Includes bibliographical references (leaves 130-145). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstracts in English and Chinese. / School code: 1307.

Finally, in the tileable bidirectional texture function (BTF), we present a modular approach to apply the BTF, which is a high dimensional texture with variable lighting and viewing directions, onto object surfaces. The basic building blocks are the BTF tiles. By constructing one set of BTF tiles, a wide variety of objects can be textured seamlessly without resynthesizing the BTF. The proposed approach can nicely decouple the surface appearance from the geometry. With this appearance geometry decoupling, one can build a library of BTF tile sets to instantaneously dress and render various objects under variable lighting and viewing conditions. The core of the proposed method is the seamless synthesis of multi-dimensional BTF tiles. To tackle the enormous data, we perform synthesis in frequency domain. This not just allows the handling of large BTF data during the synthesis, but also facilitates compact storage of the BTF in GPU memory. / For texture aware halftoning, we are tackling the lost texture details in the bitonal images generated with traditional halftoning methods. We present an optimization-based halftoning technique that preserves the texture and tone similarities between the original and the halftone images. By optimizing an objective function consisting of both the texture and the tone metrics, the generated halftone images preserve human visual sensitive details as well as the local tone. It possesses the blue-noise property and does not introduce annoying pattern. Unlike existing edge-enhancement halftoning, the proposed method does not suffer from the inability of edge detector. Our method is experimented with a various kinds of images. From the multiple experiments and the user survey, our method consistently obtains the best scores among all tested methods. / In richness-preserved screening for automated manga production, we propose a framework to generate manga-style backgrounds from real photographs. It frees manga artists from tedious and time-consuming background production. Our method divides the photo-to-manga conversion into two major process, screening and line abstraction. During the screening, our goal is to preserve the visual richness in the original photograph. The key is to exploit the pattern variety in the screening space to best maintaining the original richness in reference image in terms of texture and color. To achieve this, we select screens for different regions in the image according to tone similarity, texture similarity, and chromaticity distinguishability. Multi-dimensional scaling technique is employed for the color-to-pattern mapping. For the line abstraction, we propose a simple and effective line importance model that ranks the lines based on their geometric natures. With the line importance model, users can interactively control the level of details by tunning only a few parameters. A number of results are presented to demonstrate the effectiveness and convenience of the proposed framework. / Textures appear on most of the object surfaces or scenes around our daily lives. They are something perceptual, as they only appear when certain amount of details are perceived in a suitable scale by our eyes. In computer graphics, they are essential in enriching the perceptual appearance and improving realism in rendering. Although their nature and mechanism on how human vision system perceives and interprets textures are not well understood until now, the demand for better textural quality in rendering and imaging increases continuously. Therefore, texture related researches are becoming popular topics in the graphics community in recent years. This thesis presents several novel texture aware techniques with applications for both 3D graphics and 2D imaging. They are namely texture aware halftoning, richness-preserved screening for automated manga production, and tileable bidirectional texture function (BTF). Although these applications are from different domains in computer graphics, they are sharing a common goal of improving the texture presentation in the final outcome. Major techniques involve a texture preservation technique in halftoning process, automated manga screening method with richness preservation by pattern variety, and an effective texture tile synthesis approach for high dimensional bidirectional textures. / Pang, Wai Man. / Adviser: Heng Pheng Ann. / Source: Dissertation Abstracts International, Volume: 70-06, Section: B, page: 3609. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2008. / Includes bibliographical references (leaves 131-158). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstracts in English and Chinese. / School code: 1307.