We present a new method to perform reliable matching between different images. This method exploits a projective invariant property between concentric circles and the corresponding projected ellipses to find complete region correspondences centered on interest points. The method matches interest points allowing for a full perspective transformation and exploiting all the available luminance information in the regions. Experiments have been conducted on many different data sets to compare our approach to SIFT local descriptors. The results show the new method offers increased robustness to partial visibility, object rotation in depth, and viewpoint angle change. / Singapore-MIT Alliance (SMA)

This doctoral thesis deals with the analysis of ultrasound (US) video sequences. It specifically focuses on long-term tracking of the common carotid artery (CCA) in transversal section and measurement of its geometric parameters in a sequence of US images. The design and implementation of a system for automatic tracking of the artery is described in this thesis. The proposed system utilizes Viola-Jones detector and Hough transform to localize the artery in the image. Interest points are detected in the area of the artery wall. These points are then tracked using optical flow. The proposed system comprises a number of innovative methods which allow to perform accurate long-term measurement of parameters of CCA and store the results. A novel mathematical model describing the movement of CCA in transversal section during a cardiac cycle is defined afterwards taking the influence of breathing into consideration. A number of artificial sequences of US images based on this model have been created. These sequences were consequently used to evaluate the accuracy of the proposed system in terms of measuring the parameters of CCA. The sequences are unique because of their length which makes them suitable for evaluation of tracking accuracy even in long video sequences.

Ce mémoire de thèse traite de la mise en correspondance d'images pour des applications de vision stéréoscopique ou de stabilisation d'images de caméras vidéo. les méthodes de mise en correspondance reposent généralement sur l'utilisation de points d'intérêts dans les images, c'est-à-dire de points qui présentent de fortes discontinuités d'intensité lumineuse. Nous présentons tout d'abord un nouveau descripteur de points d'intérêt, obtenu au moyen d'un filtre anisotropique rotatif qui délivre en chaque point d'intérêt une signature mono-dimensionnelle basée sur un gradient d'intensité. Invariant à la rotationpar construction, ce descripteur possède de trés bonnes propriétés de robustesse et de discrimination. Nous proposons ensuite une nouvelle méthode d'appariement invariante aux transformations euclidiennes et affines. Cette méthode exploite la corrélation des signatures sous l'hypothèse de faibles déformations, et définit une mesure de distance nécessaire à l'appariement de points. Les résultats obtenus sur des images difficiles laissent envisager des prolongements prometteurs de cette méthode. / This thesis adresses the issue of image matching for stereoscopic vison applications and image stabilization of video cameras. Methods of mapping are generally based on the use of interest points in the images, i.e. of points which have strong discontinuities in light intensity. We first present a new descriptor of points of interest, obtained by means of an anisotropic rotary filter which delivers at each point of interest a one-dimensional signature based on an intensity gradient. Invariant to rotation by construction, thisdescriptor has very good properties of robustness and discrimination. We then propose a new matching method invariant to Euclidean and affine transformations. This method exploits the correlation of the signatures subject to moderate warping, and defines a distance measure, necesssary for the matching of points. the results obtained on difficult images augur promising extentions to this method.

Making a computer &lsquo / see&rsquo / is certainly one of the greatest challanges for today. Apart from possible applications, the solution may also shed light or at least give some idea on how, actually, the biological vision works. Many problems faced en route to successful algorithms require finding corresponding tokens in different views, which is termed the correspondence problem. For instance, given two images of the same scene from different views, if the camera positions and their internal parameters are known, it is possible to obtain the 3-Dimensional coordinates of a point in space, relative to the cameras, if the same point may be located in both images. Interestingly, the camera positions and internal parameters may be extracted solely from the images if a sufficient number of corresponding tokens can be found. In this sense, two subproblems, as the choice of the tokens and how to match these tokens, are examined. Due to the arbitrariness of the image pairs, invariant schemes for extracting and matching interest points, which were taken as the tokens to be matched, are utilised. In order to appreciate the ideas of the mentioned schemes, topics as scale-space, rotational and affine invariants are introduced. The geometry of the problem is briefly reviewed and the epipolar constraint is imposed using statistical outlier rejection methods. Despite the satisfactory matching performance of simple correlation-based matching schemes on small-baseline pairs, the simulation results show the improvements when the mentioned invariants are used on the cases for which they are strictly necessary.

This paper presents a new object detection method. The method is based on keypoints analysis and their parameters. Computed parameters are used for building a decision model using machine learning methods. The model is able to detect object in the picture based on input data and compares its similarity to the chosen example. The new method is described in detail, its accuracy is evaluated and this accuracy is compared to other existing detectors. The new method’s detection ability is by more than 40% better than detection ability of detectors like SURF. In order to understand the object detection this paper describes the process step by step including popular algorithms designed for specific roles in each step.

This master's thesis deals with the interest point detection in images. The main goal is to create an application for making panoramic photos, which is based on this detection. The application uses the SIFT detector for finding keypoints of the image. Afterwards the geometrical consistence of image points using the RANSAC algorithm is finds out and images are transform into panorama in accordance with this geometrical consistence. Finally techniques for blending of transition between photos are used.

The goal of this thesis is to get familiarized with software component Microsoft .NET Framework and the problems of methods of interest point detection. On the basis of this knowledge, it is required to develop a method for the evaluation of used detectors and then implement a console application that is simply able to evaluate the results of interest point detection methods. Such evaluation is important in the process of development of the algorithms for the detection using projected method of evaluation. The interest points are searched in several images which represent the same scene from different angles of view. The requirements also inculde creation of graphic user interface that allows an easy way to setup the evaluation conditions.

Feature extractors in Visual Odometry pipelines rarely exploit depth signals, even though depth sensors and RGB-D cameras are commonly used in later stages of Visual Odometry systems. Nonetheless, depth sensors from RGB-D cameras function even with no external light and can provide feature extractors with additional structural information otherwise invisible in RGB images. Deep Learning feature extractors, which have recently been shown to outperform their classical counterparts, still only exploit RGB information. Against this background, this thesis presents a Self-Supervised Deep Learning feature extraction algorithm that employs both RGB and depth signals as input. The proposed approach builds upon the existing deep learning feature extractors, adapting the architecture and training procedure to introduce the depth signal. The developed RGB-D system is compared with an RGB-only feature extractor in a qualitative study on keypoints’ location and a quantitative evaluation on pose estimation. The qualitative evaluation demonstrates that the proposed system exploits information from both RGB and depth domains, and it robustly adapts to the degradation of either of the two input signals. The pose estimation results indicate that the RGB-D system performs comparably to the RGB-only one in normal and low-light conditions. Thanks to the usage of depth information, the RGB-D feature extractor can still operate, showing only limited performance degradation, even in completely dark environments, where RGB methods fail due to a lack of input information. The combined qualitative and quantitative results suggest that the proposed system extracts features based on both RGB and depth input domains and can autonomously transition from normal brightness to a no-light environment, by exploiting depth signal to compensate for the degraded RGB information. / Detektering av nyckelpunkter för Visuell Odometri (VO) utnyttjar sällan information om djup i bilder, även om avståndssensorer och RGB-D-kameror ofta används i senare skeden av VO pipelinen. RGB-D-kamerors avståndsestimering fungerar även utan externt ljus. De kan förse nyckelpunktsdetektorer med ytterligare strukturell information som är svårt att extrahera enbart från RGB-bilder. Detektering av nyckelpunkter, med hjälp av Deep Learning metoder, har nyligen visat sig överträffa sina klassiska motsvarigheter som fortfarande endast utnyttjar bildinformation. Denna avhandling presenterar en algoritm för självövervakande nyckelpunktsdetektering med djupinlärning, som använder både RGB-bilder och avståndsinformation som indata. Det föreslagna tillvägagångssättet bygger på en befintlig arkitektur, som har anpassats för att också kunna hantera informationen om djupet i bilder. Den utvecklade RGB-D nyckelpunktsdetektorn har jämförts med en detektor som enbart baseras på RGB-bilder. Det har både gjorts en kvalitativ utvärdering av nyckelpunkternas läge och en kvantitativ utvärdering av detektorns förmåga på VO-tillämpningar, dvs estimering av position och orientering. Den kvalitativa utvärderingen av nyckelpunkterna visar att det föreslagna systemet kan utnyttja både information från bild- och djupdomänen. Den visar även att detektorn är robust mot försämringar av båda bilderna och djupinformationen. Evalueringen visar att den utvecklade RGB-D-metoden och en standardetektor uppnår jämförbara resultat under normala och svaga ljusförhållanden. Dock, tack vare användningen av tillgänglig djupinformation kan RGB-D-metoden fortfarande fungera i helt mörka förhållanden, med endast begränsad försämring av prestanda. I dessa scenarion misslyckas RGB-metoder på grund av brist på användbar bildinformation. De kombinerade kvalitativa och kvantitativa resultaten tyder på att det föreslagna systemet extraherar egenskaper som baseras på både bild- och djupinmatningsområden och kan självständigt övergå mellan normala och ljusfattiga förhållanden genom att utnyttja djup för att kompensera för den försämrade bildinformationen.

Scale-invariant interest points have found several highly successful applications in computer vision, in particular for image-based matching and recognition. This paper presents a theoretical analysis of the scale selection properties of a generalized framework for detecting interest points from scale-space features presented in Lindeberg (Int. J. Comput. Vis. 2010, under revision) and comprising: an enriched set of differential interest operators at a fixed scale including the Laplacian operator, the determinant of the Hessian, the new Hessian feature strength measures I and II and the rescaled level curve curvature operator, as well as an enriched set of scale selection mechanisms including scale selection based on local extrema over scale, complementary post-smoothing after the computation of non-linear differential invariants and scale selection based on weighted averaging of scale values along feature trajectories over scale. A theoretical analysis of the sensitivity to affine image deformations is presented, and it is shown that the scale estimates obtained from the determinant of the Hessian operator are affine covariant for an anisotropic Gaussian blob model. Among the other purely second-order operators, the Hessian feature strength measure I has the lowest sensitivity to non-uniform scaling transformations, followed by the Laplacian operator and the Hessian feature strength measure II. The predictions from this theoretical analysis agree with experimental results of the repeatability properties of the different interest point detectors under affine and perspective transformations of real image data. A number of less complete results are derived for the level curve curvature operator. / <p>QC 20121003</p> / Image descriptors and scale-space theory for spatial and spatio-temporal recognition

This paper deals with object detection and tracking using iterest points. Existing approaches are described here. Inovated method based on Generalized Hough transform and iterative Hough-space searching is  proposed in this paper. Generality of proposed detector is shown in various types of objects. Object tracking is designed as frame by frame detection.