Yes / Canny (IEEE Trans. Pattern Anal. Image Proc. 8(6):679-698, 1986) suggested that an optimal edge detector should maximize both signal-to-noise ratio and localization, and he derived mathematical expressions for these criteria. Based on these criteria, he claimed that the optimal step edge detector was similar to a derivative of a gaussian. However, Canny's work suffers from two problems. First, his derivation of localization criterion is incorrect. Here we provide a more accurate localization criterion and derive the optimal detector from it. Second, and more seriously, the Canny criteria yield an infinitely wide optimal edge detector. The width of the optimal detector can however be limited by considering the effect of the neighbouring edges in the image. If we do so, we find that the optimal step edge detector, according to the Canny criteria, is the derivative of an ISEF filter, proposed by Shen and Castan (Graph. Models Image Proc. 54:112-133, 1992). In addition, if we also consider detecting blurred (or non-sharp) gaussian edges of different widths, we find that the optimal blurred-edge detector is the above optimal step edge detector convolved with a gaussian. This implies that edge detection must be performed at multiple scales to cover all the blur widths in the image. We derive a simple scale selection procedure for edge detection, and demonstrate it in one and two dimensions.

<p>Shape-based matching (SBM) is a method for matching objects in greyscale images. It extracts edges from search images and matches them to a model using a similarity measure. In this thesis we extend SBM to find the tilt and height position of the object in addition to the z-plane rotation and x-y-position. The search is conducted using a scale pyramid to improve the search speed. A 3D matching can be done for small tilt angles by using SBM on height data and extending it with additional steps to calculate the tilt of the object. The full pose is useful for picking objects with an industrial robot.</p><p>The tilt of the object is calculated using a RANSAC plane estimator. After the 2D search the differences in height between all corresponding points of the model and the live image are calculated. By estimating a plane to this difference the tilt of the object can be calculated. Using the tilt the model edges are tilted in order to improve the matching at the next scale level.</p><p>The problems that arise with occlusion and missing data have been studied. Missing data and erroneous data have been thresholded manually after conducting tests where automatic filling of missing data did not noticeably improve the matching. The automatic filling could introduce new false edges and remove true ones, thus lowering the score.</p><p>Experiments have been conducted where objects have been placed at increasing tilt angles. The results show that the matching algorithm is object dependent and correct matches are almost always found for tilt angles less than 10 degrees. This is very similar to the original 2D SBM because the model edges does not change much for such small angels. For tilt angles up to about 25 degrees most objects can be matched and for nice objects correct matches can be done at large tilt angles of up to 40 degrees.</p>

A time series is composed of a sequence of data items that are measured at uniform intervals. Many application areas generate or manipulate time series, including finance, medicine, digital audio, and motion capture. Efficiently searching a large time series database is still a challenging problem, especially when partial or subseries matches are needed. This thesis proposes a new denition of subseries join, a symmetric generalization of subseries matching, which finds similar subseries in two or more time series datasets. A solution is proposed to compute the subseries join based on a hierarchical feature representation. This hierarchical feature representation is generated by an anisotropic diffusion scale-space analysis and a non-uniform segmentation method. Each segment is represented by a minimal polynomial envelope in a reduced-dimensionality space. Based on the hierarchical feature representation, all features in a dataset are indexed in an R-tree, and candidate matching features of two datasets are found by an R-tree join operation. Given candidate matching features, a dynamic programming algorithm is developed to compute the final subseries join. To improve storage efficiency, a hierarchical compression scheme is proposed to compress features. The minimal polynomial envelope representation is transformed to a Bezier spline envelope representation. The control points of each Bezier spline are then hierarchically differenced and an arithmetic coding is used to compress these differences. To empirically evaluate their effectiveness, the proposed subseries join and compression techniques are tested on various publicly available datasets. A large motion capture database is also used to verify the techniques in a real-world application. The experiments show that the proposed subseries join technique can better tolerate noise and local scaling than previous work, and the proposed compression technique can also achieve about 85% higher compression rates than previous work with the same distortion error.

A time series is composed of a sequence of data items that are measured at uniform intervals. Many application areas generate or manipulate time series, including finance, medicine, digital audio, and motion capture. Efficiently searching a large time series database is still a challenging problem, especially when partial or subseries matches are needed. This thesis proposes a new denition of subseries join, a symmetric generalization of subseries matching, which finds similar subseries in two or more time series datasets. A solution is proposed to compute the subseries join based on a hierarchical feature representation. This hierarchical feature representation is generated by an anisotropic diffusion scale-space analysis and a non-uniform segmentation method. Each segment is represented by a minimal polynomial envelope in a reduced-dimensionality space. Based on the hierarchical feature representation, all features in a dataset are indexed in an R-tree, and candidate matching features of two datasets are found by an R-tree join operation. Given candidate matching features, a dynamic programming algorithm is developed to compute the final subseries join. To improve storage efficiency, a hierarchical compression scheme is proposed to compress features. The minimal polynomial envelope representation is transformed to a Bezier spline envelope representation. The control points of each Bezier spline are then hierarchically differenced and an arithmetic coding is used to compress these differences. To empirically evaluate their effectiveness, the proposed subseries join and compression techniques are tested on various publicly available datasets. A large motion capture database is also used to verify the techniques in a real-world application. The experiments show that the proposed subseries join technique can better tolerate noise and local scaling than previous work, and the proposed compression technique can also achieve about 85% higher compression rates than previous work with the same distortion error.

A fundamental problem in spacecraft proximity operations is the determination of the 6 degree of freedom relative navigation solution between the observer reference frame and a reference frame tied to a proximal body. For the most unconstrained case, the proximal body may be uncontrolled, and the observer spacecraft has no a priori information on the body. A spacecraft in this scenario must simultaneously map the generally poorly known body being observed, and safely navigate relative to it. Simultaneous localization and mapping(SLAM)is a difficult problem which has been the focus of research in recent years. The most promising approaches extract local features in 2D or 3D measurements and track them in subsequent observations by means of matching a descriptor. These methods exist for both active sensors such as Light Detection and Ranging(LIDAR) or laser RADAR(LADAR), and passive sensors such as CCD and CMOS camera systems. This dissertation presents a method for fusing time of flight(ToF) range data inherent to scanning LIDAR systems with the passive light field measurements of optical systems, extracting features which exploit information from each sensor, and solving the unique SLAM problem inherent to spacecraft proximity operations. Scale Space analysis is extended to unstructured 3D point clouds by means of an approximation to the Laplace Beltrami operator which computes the scale space on a manifold embedded in 3D object space using Gaussian convolutions based on a geodesic distance weighting. The construction of the scale space is shown to be equivalent to both the application of the diffusion equation to the surface data, as well as the surface evolution process which results from mean curvature flow. Geometric features are localized in regions of high spatial curvature or large diffusion displacements at multiple scales. The extracted interest points are associated with a local multi-field descriptor constructed from measured data in the object space. Defining features in object space instead of image space is shown to bean important step making the simultaneous consideration of co-registered texture and the associated geometry possible. These descriptors known as Multi-Field Diffusion Flow Signatures encode the shape, and multi-texture information of local neighborhoods in textured range data. Multi-Field Diffusion Flow Signatures display utility in difficult space scenarios including high contrast and saturating lighting conditions, bland and repeating textures, as well as non-Lambertian surfaces. The effectiveness and utility of Multi-Field Multi-Scale(MFMS) Features described by Multi-Field Diffusion Flow Signatures is evaluated using real data from proximity operation experiments performed at the Land Air and Space Robotics(LASR) Laboratory at Texas A&M University.

Many laser devices acquire directly 3D objects and reconstruct their surface. Nevertheless, the final reconstructed surface is usually smoothed out as a result of the scanner internal de-noising process and the offsets between different scans. This thesis, working on results from high precision scans, adopts the somewhat extreme conservative position, not to loose or alter any raw sample throughout the whole processing pipeline, and to attempt to visualize them. Indeed, it is the only way to discover all surface imperfections (holes, offsets). Furthermore, since high precision data can capture the slightest surface variation, any smoothing and any sub-sampling can incur in the loss of textural detail.The thesis attempts to prove that one can triangulate the raw point cloud with almost no sample loss. It solves the exact visualization problem on large data sets of up to 35 million points made of 300 different scan sweeps and more. Two major problems are addressed. The first one is the orientation of the complete raw point set, an the building of a high precision mesh. The second one is the correction of the tiny scan misalignments which can cause strong high frequency aliasing and hamper completely a direct visualization.The second development of the thesis is a general low-high frequency decomposition algorithm for any point cloud. Thus classic image analysis tools, the level set tree and the MSER representations, are extended to meshes, yielding an intrinsic mesh segmentation method.The underlying mathematical development focuses on an analysis of a half dozen discrete differential operators acting on raw point clouds which have been proposed in the literature. By considering the asymptotic behavior of these operators on a smooth surface, a classification by their underlying curvature operators is obtained.This analysis leads to the development of a discrete operator consistent with the mean curvature motion (the intrinsic heat equation) defining a remarkably simple and robust numerical scale space. By this scale space all of the above mentioned problems (point set orientation, raw point set triangulation, scan merging, segmentation), usually addressed by separated techniques, are solved in a unified framework.

SANTOS, Gustavo Adolfo Campos dos. S-SWAP: scale-space based workload analysis and prediction. 2013. 99 f. Dissertação (Mestrado em ciência da computação)- Universidade Federal do Ceará, Fortaleza-CE, 2013. / Submitted by Elineudson Ribeiro (elineudsonr@gmail.com) on 2016-07-28T19:41:50Z No. of bitstreams: 1 2013_dis_gacsantos.pdf: 3910335 bytes, checksum: 15f381ec4c1d77510c3d76424bf764aa (MD5) / Approved for entry into archive by Rocilda Sales (rocilda@ufc.br) on 2016-08-01T15:38:14Z (GMT) No. of bitstreams: 1 2013_dis_gacsantos.pdf: 3910335 bytes, checksum: 15f381ec4c1d77510c3d76424bf764aa (MD5) / Made available in DSpace on 2016-08-01T15:38:14Z (GMT). No. of bitstreams: 1 2013_dis_gacsantos.pdf: 3910335 bytes, checksum: 15f381ec4c1d77510c3d76424bf764aa (MD5) Previous issue date: 2013 / This work presents a scale-space based approach to assist dynamic resource provisioning. The application of this theory makes it possible to eliminate the presence of irrelevant information from a signal that can potentially induce wrong or late decision making. Dynamic provisioning involves increasing or decreasing the amount of resources allocated to an application in response to workload changes. While monitoring both resource consumption and application-speci c metrics is fundamental in this process since the latter is of great importance to infer information about the former, dealing with these pieces of information to provision resources in dynamic environments poses a big challenge. The presence of unwanted characteristics, or noise, in a signal that represents the monitored metrics favors misleading interpretations and is known to a ect forecast models. Even though some forecast models are robust to noise, reducing its in uence may decrease training time and increase e ciency. Because a dynamic environment demands decision making and predictions on a quickly changing landscape, approximations are necessary. Thus it is important to realize how approximations give rise to limitations in the forecasting process. On the other hand, being aware of when detail is needed, and when it is not, is crucial to perform e cient dynamic forecastings. In a cloud environment, resource provisioning plays a key role for ensuring that providers adequately accomplish their obligation to customers while maximizing the utilization of the underlying infrastructure. Experiments are shown considering simulation of both reactive and proactive strategies scenarios with a real-world trace that corresponds to access rate. Results show that embodying scale-space theory in the decision making stage of dynamic provisioning strategies is very promising. It both improves workload analysis, making it more meaningful to our purposes, and lead to better predictions. / This work presents a scale-space based approach to assist dynamic resource provisioning. The application of this theory makes it possible to eliminate the presence of irrelevant information from a signal that can potentially induce wrong or late decision making. Dynamic provisioning involves increasing or decreasing the amount of resources allocated to an application in response to workload changes. While monitoring both resource consumption and application-speci c metrics is fundamental in this process since the latter is of great importance to infer information about the former, dealing with these pieces of information to provision resources in dynamic environments poses a big challenge. The presence of unwanted characteristics, or noise, in a signal that represents the monitored metrics favors misleading interpretations and is known to a ect forecast models. Even though some forecast models are robust to noise, reducing its in uence may decrease training time and increase e ciency. Because a dynamic environment demands decision making and predictions on a quickly changing landscape, approximations are necessary. Thus it is important to realize how approximations give rise to limitations in the forecasting process. On the other hand, being aware of when detail is needed, and when it is not, is crucial to perform e cient dynamic forecastings. In a cloud environment, resource provisioning plays a key role for ensuring that providers adequately accomplish their obligation to customers while maximizing the utilization of the underlying infrastructure. Experiments are shown considering simulation of both reactive and proactive strategies scenarios with a real-world trace that corresponds to access rate. Results show that embodying scale-space theory in the decision making stage of dynamic provisioning strategies is very promising. It both improves workload analysis, making it more meaningful to our purposes, and lead to better predictions.

Made available in DSpace on 2014-06-12T15:59:48Z (GMT). No. of bitstreams: 2 arquivo5591_1.pdf: 1590353 bytes, checksum: 797ebf6670d6408c7051ea1631a937a4 (MD5) license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2007 / Nos últimos anos, têm-se registrado um crescente interesse e popularização de imagens digitais, através de dispositivos tais como câmeras digitais, celulares, webcam ou filmadoras digitais. Com a grande quantidade de informação visual disponível, cresce a dificuldade do usuário em recuperar essas informações de forma precisa e eficiente. Atualmente, existem inúmeros mecanismos de busca baseados em descrições textuais ou keywords. No entanto, existem grandes dificuldades nessa abordagem, (i ) o trabalho manual requerido para notação das imagens e (ii ) a subjetividade para essa notação. Devido a essas e outras dificuldades, os mecanismos de busca baseado em keywords geram uma grande quantidade de respostas não relevantes. Nesse contexto, grandes esforços têm sido feito na área de recuperação de imagens baseados em conteúdo, de forma a tornar esse tipo de conteúdo mais acessível aos seus usuários. A proposta geral para a dissertação é desenvolver uma nova estratégia de recuperação de imagens baseada na forma, utilizando o descritor de forma Curvature Scale Space (CSS) e Mapas Auto-Organizáveis (SOM) para descrever, classificar, indexar e recuperar imagens. Essa nova abordagem possibilita a realização de consultas por similaridade levando em consideração a semelhança entre o contorno fechado dos objetos pesquisados. As características dos objetos são representados através de uma imagem multi-escalar CSS e pr´e-processados, constituindo em dados que serão usados como treinamento da rede SOM. Nesse estudo, avaliamos a acurácia e o tempo de busca através de uma base benchmark denominada Core Experiment (CE-1B). Utilizamos variações dessa base para analisar o desempenho sobre transformações geométricas de escala, rotação e translação. Os resultados obtidos mostram que a combinação do descritor CSS e SOM representa uma estratégia promissora para recuperação de imagens, com uma alta performance de tempo

Les espèces végétales, et en particulier les espèces d'arbres, forment un cadre de choix pour un processus de reconnaissance automatique basé sur l'analyse d'images. Les critères permettant de les identifier sont en effet le plus souvent des éléments morphologiques visuels, bien décrits et référencés par la botanique, qui laissent à penser qu'une reconnaissance par la forme est envisageable. Les feuilles constituent dans ce contexte les organes végétaux discriminants les plus faciles à appréhender, et sont de ce fait les plus communément employés pour ce problème qui connaît actuellement un véritable engouement. L'identification automatique pose toutefois un certain nombre de problèmes complexes, que ce soit dans le traitement des images ou dans la difficulté même de la classification en espèces, qui en font une application de pointe en reconnaissance de formes.Cette thèse place le problème de l'identification des espèces d'arbres à partir d'images de leurs feuilles dans le contexte d'une application pour smartphones destinée au grand public. Les images sur lesquelles nous travaillons sont donc potentiellement complexes et leur acquisition peu supervisée. Nous proposons alors des méthodes d'analyse d'images dédiées, permettant la segmentation et l'interprétation des feuilles d'arbres, en se basant sur une modélisation originale de leurs formes, et sur des approches basées modèles déformables. L'introduction de connaissances a priori sur la forme des objets améliore ainsi de façon significative la qualité et la robustesse de l'information extraite de l'image. Le traitement se déroulant sur l'appareil, nous avons développé ces algorithmes en prenant en compte les contraintes matérielles liées à leur utilisation.Nous introduisons également une description spécifique des formes des feuilles, inspirée par les caractéristiques déterminantes recensées dans les ouvrages botaniques. Ces différents descripteurs fournissent des informations de haut niveau qui sont fusionnées en fin de processus pour identifier les espèces, tout en permettant une interprétation sémantique intéressante dans le cadre de l'interaction avec un utilisateur néophyte. Les performances obtenues en termes de classification, sur près de 100 espèces d'arbres, se situent par ailleurs au niveau de l'état de l'art dans le domaine, et démontrent une robustesse particulière sur les images prises en environnement naturel. Enfin, nous avons intégré l'implémentation de notre système de reconnaissance dans l'application Folia pour iPhone, qui constitue une validation de nos approches et méthodes dans un cadre réel. / Plant species, and especially tree species, constitute a well adapted target for an automatic recognition process based on image analysis. The criteria that make their identification possible are indeed often morphological visual elements, which are well described and referenced by botany. This leads to think that a recognition through shape is worth considering. Leaves stand out in this context as the most accessible discriminative plant organs, and are subsequently the most often used for this problem recently receiving a particular attention. Automatic identification however gives rise to a fair amount of complex problems, linked with the processing of images, or in the difficult nature of the species classification itself, which make it an advanced application for pattern recognition.This thesis considers the problem of tree species identification from leaf images within the framework of a smartphone application intended for a non-specialist audience. The images on which we expect to work are then potentially very complex scenes and their acquisition rather unsupervised. We consequently propose dedicated methods for image analysis, in order to segment and interpret tree leaves, using an original shape modelling and deformable templates. The introduction on prior knowledge on the shape of objects enhances significatively the quality and the robustness of the information we extract from the image. All processing being carried out on the mobile device, we developed those algorithms with concern towards the material constraints of their exploitation. We also introduce a very specific description of leaf shapes, inspired by the determining characteristics listed in botanical references. These different descriptors constitute independent sources of high-level information that are fused at the end of the process to identify species, while providing the user with a possible semantic interpretation. The classification performance demonstrated over approximately 100 tree species are competitive with state-of-the-art methods of the domain, and show a particular robustness to difficult natural background images. Finally, we integrated the implementation of our recognition system into the \textbf{Folia} application for iPhone, which constitutes a validation of our approaches and methods in a real-world use.

This thesis primarily addresses the problem of automatic measurement of ore textures by image analysis in a way that is relevant to mineral processing. Specifically, it addresses the following major hypotheses: • Automatic logging of drill core by image analysis provides a feasible alternative to manual logging by geologists. • Image analysis can quantify process mineralogy by physically meaningful parameters. • Multi-scale image analysis, over a wide range of size scales, provides potential benefits to process mineralogy that are additional to those available from small-scale analysis alone, and also better retains the information content of manual logging. • Image analysis can provide physically meaningful, ore-texture-related, additive regionalised variables that can be input to geostatistical models and the definition of domains. The central focus of the thesis is the development of an automatic, multi-scale method to identify and measure objects in an image, using a specially-developed skeleton termed the morphological CWT skeleton. This skeleton is a multi-scale extension of the morphological skeleton commonly used in image analysis, and is derived from the continuous wavelet transform (CWT). Objects take the form of hierarchical segments from image segmentation based on the CWT. Only the Mexican hat, also known as the Laplacian-of-Gaussian, wavelet is used, although other wavelet shapes are possible. The natural scale of each object is defined to be the size scale at which its CWT signal (the contrast between the interior and exterior of the object) is strongest. In addition to the natural scale, the analysis automatically records the mineral composition of both the interior and exterior of each object, and shape descriptors of the object. The measurements of natural scale, mineral composition and shape are designed to relate to: • The size to which ore must be broken in order to liberate objects. • Minerals that need to be separated by physical or chemical means once objects have been liberated. • Capability to distinguish qualitatively different ore-texture types that may have different geological origins and for which different processing regimes may provide an economic benefit. Measurements are taken over size scales from three pixels to hundreds of pixels. For the major case study the pixel size is about 50 µm, but the methodology is equally applicable to photomicrographs in which the pixel size is about 4 µm. The methodology for identifying objects in images contributes to the field of scale-space image segmentation, and has advantages in performing the following actions automatically: • Finding optimal size scales in hierarchical image segmentation (natural scale). • Merging segments that are similar and spatially close together (although not necessarily touching), using the structure of the morphological CWT skeleton, thus aiding recognition of complex structures in an image. • Defining the contrast between each segment and its surrounding segments appropriately for the size scale of the segment, in a way that extends well beyond the segment boundary. For process mineralogy this contrast quantifies mineral associations at different size scales. The notion of natural scale defined in this thesis may have applications to other fields of image processing, such as mammography and cell measurements in biological microscopy. The objects identified in images are input to cluster analysis, using a finite mixture model to group the objects into object populations according to their size, composition and shape descriptors. Each image is then characterised by the abundances of different object populations that occur in it. These abundances form additive, regionalised variables that can be input into geostatistical block models. The images are themselves input to higher-level cluster analysis based on a hidden Markov model. A collection of images is divided into different ore texture types, based on differences in the abundances of the textural object populations. The ore texture types help to define geostatistical domains in an ore body. Input images for the methodology take the form of mineral maps, in which a particular mineral has been assigned to each pixel in the image prior to analysis. A method of analysing unmapped, raw colour images of ore is also outlined, as is a new model for fracture of ore. The major case study in the thesis is an analysis of approximately 1000 metres of continuously-imaged drill core from four drill holes in the Ernest Henry iron-oxide-copper-gold ore deposit (Queensland, Australia). Thirty-one texture-related variables are used to summarise the individual half-metres of drill core, and ten major ore texture types are identified. Good agreement is obtained between locations of major changes in ore type found by automatic image analysis, and those identified from manual core logging carried out by geologists. The texture-related variables are found to explain a significant amount of the variation in comminution hardness of ore within the deposit, over and above that explained by changes in abundances of the component minerals. The thesis also contributes new algorithms with wide applicability in image processing: • A fast algorithm for computing the continuous wavelet transform of a signal or image: The new algorithm is simpler in form and several times faster than the best previously-published algorithms. It consists of a single finite impulse response (FIR) filter. • A fast algorithm for computing Euclidean geodesic distance. This algorithm runs in O(1) arithmetic operations per pixel processed, which has not been achieved by any previously published algorithm. Geodesic distance is widely used in image processing, for segmentation and shape characterisation.