Object localization refers to the detection, matching and segmentation of objects in images. The localization model presented in this paper relies on deformable templates to match objects based on shape alone. The shape structure is captured by a prototype template consisting of hand-drawn edges and contours representing the object to be localized. A multistage, multiresolution algorithm is utilized to reduce the computational intensity of the search. The first stage reduces the physical search space dimensions using correlation to determine the regions of interest where a match it likely to occur. The second stage finds approximate matches between the template and target image at progressively finer resolutions, by attracting the template to salient image features using Edge Potential Fields. The third stage entails the use of evolutionary optimization to determine control point placement for a Local Weighted Mean warp, which deforms the template to fit the object boundaries. Results are presented for a number of applications, showing the successful localization of various objects. The algorithm’s invariance to rotation, scale, translation and moderate shape variation of the target objects is clearly illustrated.

No description available.

Made available in DSpace on 2014-06-11T19:22:25Z (GMT). No. of bitstreams: 0 Previous issue date: 2008-11-21Bitstream added on 2014-06-13T20:09:30Z : No. of bitstreams: 1 souza_em_dr_prud.pdf: 3668689 bytes, checksum: aa9f8f193e7b43db7e93ccc3d9a0428b (MD5) / Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) / O multicaminho é um dos fenômenos que ocorre quando o sinal proveniente do Global Navigation Satellite System (GNSS) reflete em objetos localizados nas proximidades do levantamento e chega ao receptor via múltiplos caminhos. Geralmente, o receptor GNSS recebe além do sinal direto, também o sinal refletido, o qual é atrasado em relação ao sinal direto. Conseqüentemente, as medidas de pseudodistância (PD) e fase de batimento da onda portadora são rastreadas para um sinal composto, e não para o sinal direto, causando o erro do multicaminho. Esse efeito é uma fonte de erro significativa que ainda permanece como um desafio para a pesquisa, especialmente para o posicionamento relativo estático e cinemático em aplicações de alta precisão. Diferentemente dos demais erros, o multicaminho não é atenuado quando se formam as duplas diferenças (DD) em uma linha de base curta, por ser um efeito altamente dependente do local do levantamento. Pelo contrário, os erros de multicaminho podem aumentar no processo de dupla diferenciação. Nessa pesquisa foi proposta uma metodologia, viável em termos práticos e econômicos, capaz de detectar e corrigir o efeito do multicaminho nas observações de fase da onda portadora e PD L1 e/ou L2 para aplicações estáticas e cinemáticas, quer sejam pós-processadas ou em tempo real. Essa metodologia é baseada na Análise de Multirresolução (AMR) utilizando a Transformada de Wavelets (TW). A TW é aplicada para decompor as séries temporais dos resíduos das DDs do ajustamento em componentes de freqüências baixa e alta... / GNSS-multipath is a phenomenon that occurs when the signal from Global Navigation Satellite System (GNSS) reflects on objects surrounding the survey environment and reaches the receiver antenna through multiple paths. Usually, the GNSS receiver also collects the reflected signal, which is delayed in relation to the direct one. Consequently, the pseudorange (code) and carrier phase measurements are tracked for a composed signal, and not for the direct signal, causing a multipath error. This effect is a significant error source that still remains as a challenge for the research, especially for static and kinematic relative positioning in high-precision applications. Differently from other errors sources, multipath is not attenuated when the double differences (DD) are formed in a short baseline, because this error is highly dependent upon the surrounding environment. On the contrary, multipath errors can even increase in the double differentiation process. In this research a feasible and economic methodology, able of detecting and correcting the multipath effect from the carrier phase and pseudorange L1 and/or L2 for static and kinematic applications, post-processed or in real time. This approach is based on the Multiresolution Analysis (MRA) using the Wavelet Transform (WT)... (Complete abstract click electronic access below)

With the widespread use of deformation simulations in medical applications, the realism of the force feedback has become an important issue. In order to reach real-time performance with sufficient realism the approach of adaptivity, solution of different parts of the system with different resolutions and refresh rates, has been commonly deployed. The change in accuracy resulting from the use of adaptivity, however, has been been paid scant attention in the deformation simulation field. Presentation of error metrics is rare, while more focus is given to the real-time stability. We propose an abstract pipeline to perform error analysis for different types of deformation techniques which can consider different simulation parameters. A case study is also performed using the pipeline, and the various uses of the error estimation are discussed.

The overarching goal of this project was to improve breast cancer screening protocols first by collecting small angle x-ray scattering (SAXS) images from breast biopsy tissue, and second, by applying pattern recognition techniques as a semi-automatic screen. Wavelet based features were generated from the SAXS image data. The features were supplied to a classifier, which sorted the images into distinct groups, such as normal and tumor. <p>The main problem in the project was to find a set of features that provided sufficient separation for classification into groups of normal and tumor. In the original SAXS patterns, information useful for classification was obscured. The wavelet maps allowed new scale-based information to be uncovered from each SAXS pattern. The new information was subsequently used to define features that allowed for classification. Several calculations were tested to extract useful features from the wavelet decomposition maps. The wavelet map average intensity feature was selected as the most promising feature. The wavelet map intensity feature was improved by using pre-processing to remove the high central intensities from the SAXS patterns, and by using different wavelet bases for the wavelet decomposition. <p>The investigation undertaken for this project showed very promising results. A classification rate of 100% was achieved for distinguishing between normal samples and tumor samples. The system also showed promising results when tested on unrelated MRI data. In the future, the semi-automatic pattern recognition tool developed for this project could be automated. With a larger set of data for training and testing, the tool could be improved upon and used to assist radiologists in the detection and classification of breast lesions.

Multiresolution analysis (MRA) has fairly recently become important, and even essential, to image processing and signal analysis, and is thus having a growing impact on image and signal related areas. As one of the most famous family members of the MRA, the wavelet transform (WT) has demonstrated itself in numerous successful applications in various fields, and become one of the most powerful tools in the fields of image processing and signal analysis. Due to the fact that only the scale information is supplied in WT, the applications using the wavelet transform may be limited when the absolutely-referenced frequency and phase information are required. The Stockwell transform (ST) is a recently proposed multiresolution transform that supplies the absolutely-referenced frequency and phase information. However, the ST redundantly doubles the dimension of the original data set. Because of this redundancy, use of the ST is computationally expensive and even infeasible on some large size data sets. Thus, I propose the use of the discrete orthonormal Stockwell transform (DOST), a non-redundant version of ST. This thesis will continue to implement the theoretical research on the DOST and elaborate on some of our successful applications using the DOST. We uncover the fast calculation mechanism of the DOST using an equivalent matrix form that we discovered. We also highlight applications of the DOST in image compression and image restoration, and analyze the global and local translation properties. The local nature of the DOST suggests that it could be used in many other local applications.

The overarching goal of this project was to improve breast cancer screening protocols first by collecting small angle x-ray scattering (SAXS) images from breast biopsy tissue, and second, by applying pattern recognition techniques as a semi-automatic screen. Wavelet based features were generated from the SAXS image data. The features were supplied to a classifier, which sorted the images into distinct groups, such as normal and tumor. <p>The main problem in the project was to find a set of features that provided sufficient separation for classification into groups of normal and tumor. In the original SAXS patterns, information useful for classification was obscured. The wavelet maps allowed new scale-based information to be uncovered from each SAXS pattern. The new information was subsequently used to define features that allowed for classification. Several calculations were tested to extract useful features from the wavelet decomposition maps. The wavelet map average intensity feature was selected as the most promising feature. The wavelet map intensity feature was improved by using pre-processing to remove the high central intensities from the SAXS patterns, and by using different wavelet bases for the wavelet decomposition. <p>The investigation undertaken for this project showed very promising results. A classification rate of 100% was achieved for distinguishing between normal samples and tumor samples. The system also showed promising results when tested on unrelated MRI data. In the future, the semi-automatic pattern recognition tool developed for this project could be automated. With a larger set of data for training and testing, the tool could be improved upon and used to assist radiologists in the detection and classification of breast lesions.

It is well known in image processing in general, and hence in image segmentation in particular, that computational cost increases rapidly with the number and dimensions of the images to be processed. Several fields, such as astronomy, remote sensing, and medical imaging, use very large images, which might also be 3D and/or captured at several frequency bands, all adding to the computational expense. Multiresolution analysis is one method of increasing the efficiency of the segmentation process. One multiresolution approach is the coarse-to-fine segmentation strategy, whereby the segmentation starts at a coarse resolution and is then fine-tuned during subsequent steps. Until now, the starting resolution for segmentation has been selected arbitrarily with no clear selection criteria. The research conducted for this thesis showed that starting from different resolutions for image segmentation results in different accuracies and speeds, even for images from the same dataset. An automated method for resolution selection for an input image would thus be beneficial. This thesis introduces a framework for the selection of the best resolution for image segmentation. First proposed is a measure for defining the best resolution based on user/system criteria, which offers a trade-off between accuracy and time. A learning approach is then described for the selection of the resolution, whereby extracted image features are mapped to the previously determined best resolution. In the learning process, class (i.e., resolution) distribution is imbalanced, making effective learning from the data difficult. A variant of AdaBoost, called RAMOBoost, is therefore used in this research for the learning-based selection of the best resolution for image segmentation. RAMOBoost is designed specifically for learning from imbalanced data. Two sets of features are used: Local Binary Patterns (LBP) and statistical features. Experiments conducted with four datasets using three different segmentation algorithms show that the resolutions selected through learning enable much faster segmentation than the original ones, while retaining at least the original accuracy. For three of the four datasets used, the segmentation results obtained with the proposed framework were significantly better than with the original resolution with respect to both accuracy and time.

Multiresolution analysis (MRA) has fairly recently become important, and even essential, to image processing and signal analysis, and is thus having a growing impact on image and signal related areas. As one of the most famous family members of the MRA, the wavelet transform (WT) has demonstrated itself in numerous successful applications in various fields, and become one of the most powerful tools in the fields of image processing and signal analysis. Due to the fact that only the scale information is supplied in WT, the applications using the wavelet transform may be limited when the absolutely-referenced frequency and phase information are required. The Stockwell transform (ST) is a recently proposed multiresolution transform that supplies the absolutely-referenced frequency and phase information. However, the ST redundantly doubles the dimension of the original data set. Because of this redundancy, use of the ST is computationally expensive and even infeasible on some large size data sets. Thus, I propose the use of the discrete orthonormal Stockwell transform (DOST), a non-redundant version of ST. This thesis will continue to implement the theoretical research on the DOST and elaborate on some of our successful applications using the DOST. We uncover the fast calculation mechanism of the DOST using an equivalent matrix form that we discovered. We also highlight applications of the DOST in image compression and image restoration, and analyze the global and local translation properties. The local nature of the DOST suggests that it could be used in many other local applications.

Signals on hierarchical trees can be viewed as a generalization of discrete signals of length 2^N. In this work, we extend the classic discrete Haar wavelets to a Haar-like wavelet basis that works for signals on hierarchical trees. We first construct a specific wavelet basis and give its inverse and normalized transform matrices. As analogue to the classic case, operators and wavelet generating functions are constructed for the tree structure. This leads to the definition of multiresolution analysis on a hierarchical tree. We prove the previously selected wavelet basis is an orthogonal multiresolution. Classification of all possible wavelet basis that generate an orthogonal multiresolution is then given. In attempt to find more efficient encoding and decoding algorithms, we construct a second wavelet basis and show that it is also an orthogonal multiresolution. The encoding and decoding algorithms are given and their time complexity are analyzed. In order to link change of tree structure and encoded signal, we define weighted hierarchical tree, tree cut and extension. It is then shown that a simply relation can be established without the need for global change of the transform matrix. Finally, we apply thresholding to the transform and give an upper bound of error.