Дисертації з теми "Multi-dimensional image"

Multi-dimensional digital signals have become an intertwined part of day to day life, from digital images and videos used to capture and share life experiences, to more powerful scene representations such as light field images, which open the gate to previously challenging tasks, such as post capture refocusing or eliminating visible occlusions from a scene. This dissertation delves into the world of multi-dimensional signal processing and introduces a tool of particular use for gradient based solutions of well-known signal processing problems. Specifically, a technique to reconstruct a signal from a given gradient data set is developed in the case of two dimensional (2-D), three dimensional (3-D) and four dimensional (4-D) digital signals. The reconstruction technique is multiresolution in nature, and begins by using the given gradient to generate a multi-dimensional Haar wavelet decomposition of the signals of interest, and then reconstructs the signal by Haar wavelet synthesis, performed on successive resolution levels. The challenges in developing this technique are non-trivial and are brought about by the applications at hand. For example, in video content replacement, the gradient data from which a video sequence needs to be reconstructed is a combination of gradient values that belong to different video sequences. In most cases, such operations disrupt the conservative nature of the gradient data set. The effects of the non-conservative nature of the newly generated gradient data set are attenuated by using an iterative Poisson solver at each resolution level during the reconstruction. A second and more important challenge is brought about by the increase in signal dimensionality. In a previous approach, an intermediate extended signal with symmetric region of support is obtained, and the signal of interest is extracted from it. This approach is reasonable in 2-D, but becomes less appealing as the signal dimensionality increases. To avoid generating data that is then discarded, a new approach is proposed, in which signal extension is no longer performed. Instead, different procedures are suggested to generate a non-symmetric Haar wavelet decomposition of the signals of interest. In the case of 2-D and 3-D signals, ways to obtain this decomposition exactly from the given gradient data and the average value of the signal are proposed. In addition, ways to approximate a subset of decomposition coefficients are introduced and the visual consequences of such approximations are studied in the special case of 2-D digital images. Several ways to approximate the same subset of decomposition coefficients are developed in the special case of 4-D light field images. Experiments run on various 2-D, 3-D and 4-D test signals are included to provide an insight on the performance of the reconstruction technique. The value of the multi-dimensional reconstruction technique is then demonstrated by including it in a number of signal processing applications. First, an efficient algorithm is developed with the purpose of combining information from the gradient of a set of 2-D images with different regions in focus or different exposure times, with the purpose of generating an all-in-focus image or revealing details that were lost due to improper exposure setting. Moving on to 3-D signal processing applications, two video editing problems are studied and gradient based solutions are presented. In the first one, the objective is to seamlessly place content from one video sequence in another, while in the second one, to combine elements from two video sequences and generate a transparency effect. Lastly, a gradient based technique for editing 4-D scene representations (light fields) is presented, as well as a technique to combine information from two light fields with the purpose of generating a light field with more details of the imaged scene. All these applications show that the developed technique is a reliable tool for gradient domain based solutions of signal processing problems.
Graduate

博士
國立臺北科技大學
機電科技研究所
105
Sharp surface edges present very crucial visual information which corresponds to discontinuities in the geometry and physical properties of objects, such as major changes in surface reflectance, illumination, orientation and depth. The research in this dissertation presents a new method employing structured illumination imaging and image fusion to overcome one of the greatest difficulties in 3-D optical measurement: accurate surface sharp edges reconstruction. The surface sharp edges are extremely critical for extracting key geometric dimension information required for fulfilling the tremendous demands of various kinds of precision engineering. Measurement and reconstruction of surface sharp edges using optical surface profilometry is rather difficult, since there exist complicated optical phenomena around surface edges, such as complex diffraction or interference. Noisy reconstructed data around surface edges are commonly observed in many optical detection methods, such as optical moiré projection, confocal microscopy or interferometry. Thus, in this research, a new method is proposed to reconstruct an edge surface profile by integrating a 2-D surface edge with its neighboring 3-D surface profile. The 2-D surface edge is extracted from a high spatial resolution 2-D image, which is obtained using structured illumination imaging to achieve optical super-resolution, so the projected edge contour of the 2-D contour along the optical imaging axis can be accurately determined. The 3-D surface profile neighboring the edge is simultaneously detected by an adequate 3-D optical measuring method, such as phase-shifting based moiré fringe projection, as well as either confocal measurement or optical interferometry. Moreover, in order to have accurate 3-D surface sharp edges reconstruction, the invalid point cloud data existing around the sharp edges is filtered away from the 3-D optical detection result. The neighboring surface between the 2-D accurate detected edge and the identified 3-D surface contour is then reconstructed using extrapolation by NURBS surface fitting to detect the intersecting edge. Some experiments are performed to verify the feasibility, effectiveness and accuracy of the developed method by comparing the reconstructed 3-D sharp edge with one determined using a pre-calibrated high precision instrument. The proposed method ensures that the maximum deviation between tactile CMM measurement and our developed method can keep the measured error to be within 3

碩士
亞洲大學
財務金融學系碩士在職專班
105
This study managed to test and verify the multi-dimensional color barcode and tracking system applied by Taiwan Zhongchian Technology Co., Ltd. This business wants contribute its energy and resources to revitalize Taiwanese agriculture and food industries while safeguarding public health and improving its corporate image and performance as a gatekeeper of food products. This study examined the multi-dimensional color barcode and tracking system developed by the company for the sake of maintaining food safety. This system can help address two major problems undermining food safety. First, given the uniqueness of the multi-dimensional color barcodes, they cannot be replicated, and thus can avoid barcode fraud. In addition, the tracking system allows consumers to check the whole production chain, from the field to the table, themselves. As such, food quality issues can be resolved as completely as possible. Furthermore, the designability and scalability of the multi-dimensional color barcodes allow the barcodes to be made into different shapes, as well as patterns, and even be presented in animation, to enhance businesses’ corporate image. Since the multi-dimensional color barcode and tracking system can be developed into a database, complete information can be offered to consumers, which will, in turn, increase the efficiency of businesses’ production process and enhance their corporate performance, as a result.

De plus en plus souvent, les études médicales utilisent simultanément de multiples modalités d'acquisition d'image, produisant ainsi des données multidimensionnelles comportant beaucoup d'information supplémentaire dont l'interprétation et le traitement deviennent délicat. Par exemple, les études sur l'ischémie cérébrale se basant sur la combinaison de plusieurs images IRM, provenant de différentes séquences d'acquisition, pour prédire l'évolution de la zone nécrosée, donnent de bien meilleurs résultats que celles basées sur une seule image. Ces approches nécessitent cependant l'utilisation d'algorithmes plus complexes pour réaliser les opérations de filtrage, segmentation et de clustering. Une approche robuste pour répondre à ces problèmes de traitements de données multidimensionnelles est le Mean Shift qui est basé sur l'analyse de l'espace des caractéristiques et l'estimation non-paramétrique par noyau de la densité de probabilité. Dans cette thèse, nous étudions les paramètres qui influencent les résultats du Mean Shift et nous cherchons à optimiser leur choix. Nous examinons notamment l'effet du bruit et du flou dans l'espace des caractéristiques et comment le Mean Shift doit être paramétrés pour être optimal pour le débruitage et la réduction du flou. Le grand succès du Mean Shift est principalement du au réglage intuitif de ces paramètres de la méthode. Ils représentent l'échelle à laquelle le Mean Shift analyse chacune des caractéristiques. En se basant sur la méthode du Plug In (PI) monodimensionnel, fréquemment utilisé pour le filtrage Mean Shift et permettant, dans le cadre de l'estimation non-paramétrique par noyau, d'approximer le paramètre d'échelle optimal, nous proposons l'utilisation du PI multidimensionnel pour le filtrage Mean Shift. Nous évaluons l'intérêt des matrices d'échelle diagonales et pleines calculées à partir des règles du PI sur des images de synthèses et naturelles. Enfin, nous proposons une méthode de segmentation automatique et volumique combinant le filtrage Mean Shift et la croissance de région ainsi qu'une optimisation basée sur les cartes de probabilité. Cette approche est d'abord étudiée sur des images IRM synthétisées. Des tests sur des données réelles issues d'études sur l'ischémie cérébrale chez le rats et l'humain sont aussi conduits pour déterminer l'efficacité de l'approche à prédire l'évolution de la zone de pénombre plusieurs jours après l'accident vasculaire et ce, à partir des IRM réalisées peu de temps après la survenue de cet accident. Par rapport aux segmentations manuelles réalisées des experts médicaux plusieurs jours après l'accident, les résultats obtenus par notre approche sont mitigés. Alors qu'une segmentation parfaite conduirait à un coefficient DICE de 1, le coefficient est de 0.8 pour l'étude chez le rat et de 0.53 pour l'étude sur l'homme. Toujours en utilisant le coefficient DICE, nous déterminons la combinaison de d'images IRM conduisant à la meilleure prédiction
Medical studies increasingly use multi-modality imaging, producing multidimensional data that bring additional information that are also challenging to process and interpret. As an example, for predicting salvageable tissue, ischemic studies in which combinations of different multiple MRI imaging modalities (DWI, PWI) are used produced more conclusive results than studies made using a single modality. However, the multi-modality approach necessitates the use of more advanced algorithms to perform otherwise regular image processing tasks such as filtering, segmentation and clustering. A robust method for addressing the problems associated with processing data obtained from multi-modality imaging is Mean Shift which is based on feature space analysis and on non-parametric kernel density estimation and can be used for multi-dimensional filtering, segmentation and clustering. In this thesis, we sought to optimize the mean shift process by analyzing the factors that influence it and optimizing its parameters. We examine the effect of noise in processing the feature space and how Mean Shift can be tuned for optimal de-noising and also to reduce blurring. The large success of Mean Shift is mainly due to the intuitive tuning of bandwidth parameters which describe the scale at which features are analyzed. Based on univariate Plug-In (PI) bandwidth selectors of kernel density estimation, we propose the bandwidth matrix estimation method based on multi-variate PI for Mean Shift filtering. We study the interest of using diagonal and full bandwidth matrix with experiment on synthesized and natural images. We propose a new and automatic volume-based segmentation framework which combines Mean Shift filtering and Region Growing segmentation as well as Probability Map optimization. The framework is developed using synthesized MRI images as test data and yielded a perfect segmentation with DICE similarity measurement values reaching the highest value of 1. Testing is then extended to real MRI data obtained from animals and patients with the aim of predicting the evolution of the ischemic penumbra several days following the onset of ischemia using only information obtained from the very first scan. The results obtained are an average DICE of 0.8 for the animal MRI image scans and 0.53 for the patients MRI image scans; the reference images for both cases are manually segmented by a team of expert medical staff. In addition, the most relevant combination of parameters for the MRI modalities is determined

The advancement of medical imaging techniques such as fundus photography and breast magnetic resonance imaging (MRI) has shown tremendous improvement in the quality of multidimensional image produced. The image segmentation technology is used to partition the medical image into different regions for accurate identification and segregation of diseased area. Hence, the medical image is a vital entity to diagnose several pathological conditions. However, Multidimensional medical image analysis with automatic segmentation techniques these medical images have problems such as: 1. lack inherent spatial resolution; 2. contains different form of noise; 3. have boundary with the similar color intensity; and 4. populated with non-uniform illumination across the image and other imaging ambiguities. In many clinical studies, the segmentation process can be carried out either manually or automatically. Manual segmentation for the identification of several landmarks in medical images has been popularly considered, but is time consuming, tedious, error prone and observer-dependent. On the other hand, automatic segmentation technique are highly desirable because of its robustness, improved efficiency, reliability and faster computation. Therefore, the development of an automatic segmentation technique for the medical images has become an integral part of the medical diagnosis system that yields a practical insight. However, achieving a desirable result from automatic segmentation is still challenging. This is because; variation is seen in image features for different cases, even when produced with same imaging technique. The broad aim of this thesis is to identify the robust and automatic segmentation technique overcoming the issues seen in medical images and hence can assist doctors for the evaluation and detection of several pathologies. The objective is fulfilled by developing automatic segmentation algorithms and provide solutions to tackle challenges associated in two different imaging modalities: fundus photography (2D) and breast MRI (3D). The result is a series of work associated with the problem identification, analysis and a desirable solution with qualitative and quantitative validation. Specifically, we have strengthened the state-of-the-art by making the following novel contributions: 1. The analysis of retinal blood vessel is crucial for finding several pathological disorder that manifest through human eye. Therefore, blood vessel segmentation in fundus photography has great importance in medical image analysis. From the experiment, we observed that the retinal images with lesions, exudate’s, non-uniformed illuminations and pathological artefacts have intrinsic problems such as the absence of thin vessels and detection of false vessels. In our work, we developed an automatic blood vessel segmentation framework, which is effective in analysing retinal blood vessels on noisy, pathological and abnormal retinal images. Initially, the noise is minimized with image subtraction technique using morphological operation. Then, we investigated thin and thick blood vessels separately. Thin vessels are detected using local phase-preserving denoising, line detection, local normalization, and maximum entropy thresholding. Local phase-preservation denoising removes the additional noise while preserving phase information (detailed) of the image. Thick vessels are segmented using maximum entropy thresholding. The performance of the proposed methods is carried in four popular databases (DRIVE, STARE, CHASE DB1, HRF). The result shows that the proposed segmentation method is automatic, accurate and computationally efficient. Furthermore, the proposed methods is found to be superior when compared with the other methods in the state of art. 2. The automatic optic disc (OD) segmentation is a challenging task for the images, which are under the influence of noise, uneven illumination and pathologies. As per the state-of-art, development of OD segmentation is still a challenging task because of several reasons such as 1) Ophthalmic pathologies causes the change of color, shape or depth of OD 2) Retinal pathologies (exudate, lesion), sometimes possess similar properties causing a false identification of OD. 3) Different factors like illuminations and contrast irregularities, boundary artefacts and blurred image edges makes segmentation complicated and requires pixel to pixel analysis. 4) Also the texture feature of OD vary for different images, adding more challenges, thus requiring a pre-processing step prior to the segmentation. 5) If the vessels are dense and around OD, the identification the OD boundary becomes difficult. To solve the above-mentioned challenges, a new method for the accurate localization and detection of the optic disc is developed. The process utilizes kmeans clustering over foreground and background estimated images to obtain the brightest cluster. The obtained results are merged together to estimate the OD center. The OD boundary is then estimated using circular Hough transform (CHT) using the radius and center obtained in the initial step. The boundary estimation is also obtained from superpixels method. Finally, the OD boundary pixels are identified with the geometrical model over the edge information obtained from superpixels and CHT. The experiments carried out on seven publicly available database verify the efficiency of proposed methods. In addition, the outstanding results while compared with the other proposed methods in the current state of art proves the superiority of proposed methods. 3. A novel and accurate segmentation method of the breast region of interest (BROI) and breast density (BD) in breast MRI is proposed. The precise segmentation of BROI and BD is challenging, especially in noisy magnetic resonance images (MRI) due to similar intensity levels and the closely connected boundaries between BROI and other anatomical structure such as heart, lung and pectoral muscle. The segmentation of BROI is carried out in three major steps. Initially, we utilize adaptive wiener filtering and k-means clustering to denoised image by preserving edges and unwanted artefacts. Then, active contour based level sets is used to eliminate the heart area from the denoised image. Initial contour points for the active contour methods are determined by the maximum entropy thresholding and convolution method. Finally, a pectoral muscle is removed to obtain a BROI segmentation by using a morphological operations and local adaptive thresholding methods. The segmentation of BD is obtained with 4 level fuzzy c-means (FCM) thresholding methods on the result image obtained from BROI segmentation. The validation of proposed methods is performed using the 1350 breast images from 15 female subjects. The obtained result show that the proposed method is automatic, fast and efficient. 4. The segmentation of breast lesions in breast MRI is considered as a important and challenging task in medical image analysis. Noise, intensity similarity of lesions and other tissues, and variable shape and size of lesion are the primary challenges during the process of lesion segmentation. Hence, the framework for the accurate segmentation of breast lesion from the DCE MRI image is proposed. The framework is built using max flow and min cut problems in the continuous domain over the denoised image. The proposed method is achieved in three steps. Firstly, in the pre-processing step, the post contrast and pre-contrast image are subtracted. This is followed by image registration that benefits by enhancing the tumor area. Secondly, a phase preservation denoising and pixel-wise adaptive Wiener filtering technique are used which is followed by max flow and min cut problems in the continuous domain. A denoising mechanism clears the noise in the image by preserving the useful and detailed features such as edges. Then, a tumor detection is done using continuous max flow. Finally, morphological operation is used as a post-processing step to further delineate the obtained results. The efficiency of the proposed method is verified with the series of qualitative and quantitative experiments carried out on 21 cases with two different MR image resolution. The obtained results when compared with the manually segmented results demonstrates the quality of segmentation obtained from the proposed method. The segmentation experiments for all above-mentioned four proposed algorithms are performed on Matlab R2013b running under Intel(R) core(TM) i5-4570s CPU@ 2.90 Ghz with 8GB of RAM. In an effort to test the performance of the proposed algorithms, both the public and private datasets with the manually drawn ground truth image are used. Moreover, the qualitative and quantitative measurements were used as a way to verify the robustness of the proposed algorithms. Also, the result were compared with the recent state-of-art which demonstrate the enhanced performance and advancement of the proposed methods. Finally, our overall results on the proposed methods show that the proposed algorithms are automatic, accurate and computationally efficient.

Indiana University-Purdue University Indianapolis (IUPUI)
We present a novel, variational and statistical approach for model-based segmentation. Our model generalizes the Chan-Vese model, proposed for concurrent segmentation of multiple objects embedded in the same image domain. We also propose a novel shape descriptor, namely the Multi-Compartment Distance Functions or mcdf. Our proposed framework for segmentation is two-fold: first, several training samples distributed across various classes are registered onto a common frame of reference; then, we use a variational method similar to Active Shape Models (or ASMs) to generate an average shape model and hence use the latter to partition new images. The key advantages of such a framework is: (i) landmark-free automated shape training; (ii) strict shape constrained model to fit test data. Our model can naturally deal with shapes of arbitrary dimension and topology(closed/open curves). We term our model Active Geometric Model, since it focuses on segmentation of geometric shapes. We demonstrate the power of the proposed framework in two important medical applications: one for morphology estimation of 3D Motor Neuron compartments, another for thickness estimation of Henle's Fiber Layer in the retina. We also compare the qualitative and quantitative performance of our method with that of several other state-of-the-art segmentation methods.

碩士
國立成功大學
資訊工程研究所
88
Low back pain is one of the most frequent problems treated by orthopedic surgeons. Four out of five adults will experience significant low back pain sometime during their lifetime. For better understanding of this disease, we need to know how the lower of the lumbar spine gets infected histologically. First, the region is on the fourth and fifth area of the lumbar spine that connects the upper body to the lower body. The important part of spine provides mechanism for turning, twisting or bending; and also standing. In other words, it is the one that provides you with both mobility and strength. It is obvious that people will be uncomfortable when getting pathological changes at this part. And low back pain is the one. What are the common causes? Besides the skeletal-muscular problem, most reasons are pain induced to the nerve root, which has been compressed by the foramen cramped. The foramen narrows when the facet joint aged, ligament hypertrophy or protruding disk. Following the clinical pass way, MR imaging will be the only diagnosis and measurement tool by its limited information. Maybe CT images will be the two-dimension reference, but few people try to register the images of the both modality and produce the three-dimension matching model. Our research tries to register the two images and get more thorough information for the disease diagnosis. As the segmentation of the MR bone model is a very tough job, we develop a new method to register the image by optimizing the gradient information accumulated around the bony boundary areas with respect to a 3D reference model obtained from CT images. The system consists of all the necessary pre-processing, 2D segmentation, 3D reconstruction, fusion, and rendering sub-system for multi-modality spinal images to perform the desired operations and verify the robustness and accuracy for the algorithm. This system will be useful to observe the foramen and the nerve root. Since the registration can be performed without external markers, it could be a better choice for clinical usage in the diagnosis of lumbar spine.

碩士
國立臺北科技大學
機電整合研究所
93
An apparatus and method that uses six digital cameras to capture images of the foot, can reconstruct the 3D model of the foot rapidly. Users can only wear elastic socks, which have multicolored stripes or others that we design especially. A user stands on a tempered glass. First from two parallel cameras under the foot, we can compute 3D model about the sole of the foot. Other four cameras are around the upper part of the foot. Second, the adjacent images combine four curved surfaces. Finally, we combine those curved surfaces to reconstruct complete 3D feet model.

碩士
國立臺北科技大學
電機工程系研究所
95
A novel study of feature extraction technique for hyperspectral images of remote sensing is proposed. The method is based on the greedy modular eigenspace (GME) scheme, which was designed to extract the simplest and the most efficient feature modules for high-dimensional datasets. It presents a framework which consists of two algorithms, referred to as multi-dimensional correlation matrix feature extraction (MD-CMFE) and the feature scale uniformity transformation (FSUT). The MD-CMFE scheme, also known as the complete modular eigenspace (CME), can improve the performance of GME feature extraction optimally by modifying the conventional correlation coefficient operations. It is designed to extract features by a new defined three dimensional correlation matrix (3D-CM) to optimize the modular eigenspace, while FSUT is performed to fuse most correlated features from different spectrums associated with different data sources. In this paper, we also present a parallel computing technique for the feature extraction of hyperspectral images. The proposed parallel CME (PCME) scheme is introduced to reduce the computational load of CME feature extraction using the parallel computing technique. It is implemented by parallel virtual machine (PVM) to solve the huge matrix problems of CME feature extraction. The performance of the proposed method is evaluated by applying to hyperspectral images of MODIS/ASTER (MASTER) airborne simulator during the Pacrim II project. The experiments demonstrate the proposed MD-CMFE/FSUT and PCME approach is an effective scheme not only for the feature extraction but also for the band selection of high-dimensional datasets. It can improve the precision of hyperspectral image classification compared to conventional multispectral classification schemes.

碩士
國立中央大學
機械工程學系
105
This study develops a 3-dimensional object model reconstruction system. By using color stripe structure light pattern and a Realsense™ camera to capture the object surface point cloud, and then to rebuild the 3-dimensional object model. The system uses a calibration plateau as reference, by capturing their 2-D images, calculates the internal and external parameters of the camera, and rebuild a 3-D imaging plane. Projecting the encoded color-stripe pattern structure light on the object, we use two cameras and multi-view in the capturing process. Through image recognition system and stripe indexing system, we calculate the 3-D point cloud of the object surface by these 2-D image data. We also use Realsense™ 3-D camera to capture the surface point cloud, and combine multiple sets of point clouds by iterative closed point algorithm. Rebuild the 3-D object mesh, and use origin object color images to apply surface texture. We use several objects as example, applying the 3-Dimension object model reconstruction system. Capture the object images and rebuild 3-D model, and compare with Rough Models to verify the accuracy of the study.

Tao, Chenjun.
Thesis (M.Phil.)--Chinese University of Hong Kong, 2008.
Includes bibliographical references (leaves 89-97).
Abstracts in English and Chinese.
Abstract --- p.i
Acknowledgements --- p.iii
Chapter 1 --- Introduction --- p.1
Chapter 1.1 --- Window-oriented Retargeting --- p.1
Chapter 1.2 --- Abstraction Rendering --- p.4
Chapter 1.3 --- Thesis Outline --- p.6
Chapter 2 --- Related Work --- p.7
Chapter 2.1 --- Video Migration --- p.8
Chapter 2.2 --- Video Synopsis --- p.9
Chapter 2.3 --- Periodic Motion --- p.14
Chapter 2.4 --- Video Tracking --- p.14
Chapter 2.5 --- Video Stabilization --- p.15
Chapter 2.6 --- Video Completion --- p.20
Chapter 3 --- Active Window Oriented Video Retargeting --- p.21
Chapter 3.1 --- System Model --- p.21
Chapter 3.1.1 --- Foreground Extraction --- p.23
Chapter 3.1.2 --- Optimizing Active Windows --- p.27
Chapter 3.1.3 --- Initialization --- p.29
Chapter 3.2 --- Experiments --- p.32
Chapter 3.3 --- Summary --- p.37
Chapter 4 --- Multi-Style Abstract Image Rendering --- p.39
Chapter 4.1 --- Abstract Images --- p.39
Chapter 4.2 --- Multi-Style Abstract Image Rendering --- p.42
Chapter 4.2.1 --- Multi-style Processing --- p.45
Chapter 4.2.2 --- Layer-based Rendering --- p.46
Chapter 4.2.3 --- Abstraction --- p.47
Chapter 4.3 --- Experimental Results --- p.49
Chapter 4.4 --- Summary --- p.56
Chapter 5 --- Interactive Abstract Videos --- p.58
Chapter 5.1 --- Abstract Videos --- p.58
Chapter 5.2 --- Multi-Style Abstract Video --- p.59
Chapter 5.2.1 --- Abstract Images --- p.60
Chapter 5.2.2 --- Video Morphing --- p.65
Chapter 5.2.3 --- Interactive System --- p.69
Chapter 5.3 --- Interactive Videos --- p.76
Chapter 5.4 --- Summary --- p.77
Chapter 6 --- Conclusions --- p.81
Chapter A --- List of Publications --- p.83
Chapter B --- Optical flow --- p.84
Chapter C --- Belief Propagation --- p.86
Bibliography --- p.89

碩士
國立臺北科技大學
機電整合研究所
94
This study developed an apparatus and method that uses at least six digital cameras to capture images of the foot, which can reconstruct the 3D model of the foot rapidly. Users can only wear elastic socks, which have specially coded multicolored stripes or spots, and stands on the transparent plate of a platform. First from two parallel cameras under the transparent plate, the 3D contour about the sole of the foot can be computed and reconstructed through stereosis algorithm. Similarly, from the other four cameras at least around the upper part of the foot, four or more curved surfaces are combined to reconstruct 3D contour of the upper portion of the foot. Combining the bottom contours and upper contours of the foot the complete 3D foot model can be obtained.

Performance analysis often considers numerous factors contributing to performance, and the relative importance of these factors is evolving based on dynamic conditions and requirements. Investigating large numbers of factors and understanding individual factors' predictability within the ultimate performance are challenging tasks. A visual analytics approach that integrates interactive analysis, novel visual representations, and predictive machine learning models can provide new capabilities to examine performance effectively and thoroughly. Currently, only limited research has been done on the possible applications of visual analytics for performance evaluation. In this dissertation, two specific types of performance analysis are presented: (1) organizational employee performance evaluation and (2) performance improvement of machine learning models with interactive feature selection. Both application scenarios leverage the human-in-the-loop approach to assist the identification of influential factors. For organizational employee performance evaluation, a novel visual analytics system, MetricsVis, is developed to support exploratory organizational performance analysis. MetricsVis incorporates hybrid evaluation metrics that integrate quantitative measurements of observed employee achievements and subjective feedback on the relative importance of these achievements to demonstrate employee performance at and between multiple levels regarding the organizational hierarchy. MetricsVis II extends the original system by including actual supervisor ratings and user-guided rankings to capture preferences from users through derived weights. Comparing user preferences with objective employee workload data enables users to relate user evaluation to historical observations and even discover potential bias. For interactive feature selection and model evaluation, a visual analytics system, FeatureExplorer, allows users to refine and diagnose a model iteratively by selecting features based on their domain knowledge, interchangeable features, feature importance, and the resulting model performance. FeatureExplorer enables users to identify stable, trustable, and credible predictive features that contribute significantly to a prediction model.