The work in this thesis aims at studying the problems related to the robustness of a face recognition system where specific attention is given to the issues of handling the image variation complexity and inherent limited Unique Characteristic Information (UCI) within the scope of unfamiliar identity recognition environment. These issues will be the main themes in developing a mutual understanding of extraction and classification tasking strategies and are carried out as a two interdependent but related blocks of research work. Naturally, the complexity of the image variation problem is built up from factors including the viewing geometry, illumination, occlusion and other kind of intrinsic and extrinsic image variation. Ideally, the recognition performance will be increased whenever the variation is reduced and/or the UCI is increased. However, the variation reduction on 2D facial images may result in loss of important clues or UCI data for a particular face alternatively increasing the UCI may also increase the image variation. To reduce the lost of information, while reducing or compensating the variation complexity, a hybrid technique is proposed in this thesis. The technique is derived from three conventional approaches for the variation compensation and feature extraction tasks. In this first research block, transformation, modelling and compensation approaches are combined to deal with the variation complexity. The ultimate aim of this combination is to represent (transformation) the UCI without losing the important features by modelling and discard (compensation) and reduce the level of the variation complexity of a given face image. Experimental results have shown that discarding a certain obvious variation will enhance the desired information rather than sceptical in losing the interested UCI. The modelling and compensation stages will benefit both variation reduction and UCI enhancement. Colour, gray level and edge image information are used to manipulate the UCI which involve the analysis on the skin colour, facial texture and features measurement respectively. The Derivative Linear Binary transformation (DLBT) technique is proposed for the features measurement consistency. Prior knowledge of input image with symmetrical properties, the informative region and consistency of some features will be fully utilized in preserving the UCI feature information. As a result, the similarity and dissimilarity representation for identity parameters or classes are obtained from the selected UCI representation which involves the derivative features size and distance measurement, facial texture and skin colour. These are mainly used to accommodate the strategy of unfamiliar identity classification in the second block of the research work. Since all faces share similar structure, classification technique should be able to increase the similarities within the class while increase the dissimilarity between the classes. Furthermore, a smaller class will result on less burden on the identification or recognition processes. The proposed method or collateral classification strategy of identity representation introduced in this thesis is by manipulating the availability of the collateral UCI for classifying the identity parameters of regional appearance, gender and age classes. In this regard, the registration of collateral UCI s have been made in such a way to collect more identity information. As a result, the performance of unfamiliar identity recognition positively is upgraded with respect to the special UCI for the class recognition and possibly with the small size of the class. The experiment was done using data from our developed database and open database comprising three different regional appearances, two different age groups and two different genders and is incorporated with pose and illumination image variations.

In this paper we introduce a novel representation of the significant changes in curvature along the bounding contour of planar shape. We call the representation the curvature primal sketch. We describe an implemented algorithm that computes the curvature primal sketch and illustrate its performance on a set of tool shapes. The curvature primal sketch derives its name from the close analogy to the primal sketch representation advocated by Marr for describing significant intensity changes. We define a set of primitive parameterized curvature discontinuities, and derive expressions for their convolutions with the first and second derivatives of a Gaussian. The convolved primitives, sorted according to the scale at which they are detected, provide us with a multi-scaled interpretation of the contour of a shape.

No description available.

With the rapid development of computer and information technologies, medical imaging has become one of the major sources of information for therapy and research in medicine, biology and other fields. Along with the advancement of medical imaging techniques, computer-aided detection and diagnosis (CAD/CADx) has recently emerged to become one of the major research subjects within the area of diagnostic radiology and medical image analysis. This thesis presents two multi-level learning-based approaches for medical image understanding with applications of CAD/CADx. The so-called "multi-level learning strategy" relies on that supervised and unsupervised statistical learning techniques are utilized to hierarchically model and analyze the medical image content in a "bottom up" way. As the first approach, a learning-based algorithm for automatic medical image classification based on sparse aggregation of learned local appearance cues is proposed. The algorithm starts with a number of landmark detectors to collect local appearance cues throughout the image, which are subsequently verified by a group of learned sparse spatial configuration models. In most cases, a decision could already be made at this stage by simply aggregating the verified detections. For the remaining cases, an additional global appearance filtering step is employed to provide complementary information to make the final decision. This approach is evaluated on a large-scale chest radiograph view identification task and a multi-class radiograph annotation task, demonstrating its improved performance in comparison with other state-of-the-art algorithms. It also achieves high accuracy and robustness against images with severe diseases, imaging artifacts, occlusion, or missing data. As the second approach, a learning-based approach for automatic segmentation of ill-defined and spiculated mammographic masses is presented. The algorithm starts with statistical modeling of exemplar-based image patches. Then, the segmentation problem is regarded as a pixel-wise labeling problem on the produced mass class-conditional probability image, where mass candidates and clutters are extracted. A multi-scale steerable ridge detection algorithm is further employed to detect spiculations. Finally, a graph-cuts technique is employed to unify all outputs from previous steps to generate the final segmentation mask. The proposed method specifically tackles the challenge of inclusion of mass margin and associated extension for segmentation, which is considered to be a very difficult task for many conventional methods. / Master of Science

This thesis is concerned with the problem of place recognition for mobile robots. How can a robot determine its location from an image or sequence of images, without any prior knowledge of its position, even in a world where many places look identical? We outline a new probabilistic approach to the problem, which we call Fast Appearance Based Mapping or FAB-MAP. Our map of the environment consists of a set of discrete locations, each with an associated appearance model. For every observation collected by the robot, we compute a probability distribution over the map, and either create a new location or update our belief about the appearance of an existing location. The technique can be seen as a new type of SLAM algorithm, where the appearance of locations (rather than their position) is subject to estimation. Unlike existing SLAM systems, our appearance based technique does not rely on keeping track of the robot in any metric coordinate system. Thus it is applicable even when informative observations are available only intermittently. Solutions to the loop closure detection problem, the kidnapped robot problem and the multi-session mapping problem arise as special cases of our general approach. Abstract Our probabilistic model introduces several technical advances. The model incorporates correlations between visual features in a novel way, which is shown to improve system performance. Additionally, we explicitly compute an approximation to the partition function in our Bayesian formulation, which provides a natural probabilistic measure of when a new observation should be assigned to a location not already present in the map. The technique is applicable even in visually repetitive environments where many places look the same. Abstract Finally, we define two distinct approximate inference procedures for the model. The first of these is based on concentration inequalities and has general applicability beyond the problem considered in this thesis. The second approach, built on inverted index techniques, is tailored to our specific problem of place recognition, but achieves extreme efficiency, allowing us to apply FAB-MAP to navigation problems on the largest scale. The thesis concludes with a visual SLAM experiment on a trajectory 1,000 km long. The system successfully detects loop closures with close to 100% precision and requires average inference time of only 25 ms by the end of the trajectory.

Cancer is one of the leading causes of premature deaths across the world today, and there is an urgent need for imaging techniques that can help in early diagnosis and treatment planning for cancer patients. In the last four decades, magnetic resonance imaging (MRI) has emerged as one of the leading modalities for non-invasive imaging of tumours. By using dynamic contrast-enhanced magnetic resonance imaging (DCEMRI), this modality can be used to acquire information about perfusion and vascularity of tumours, which can help in predicting response to treatment. There are many factors that complicate the analysis of DCE-MRI data, and make clinical predictions based on it unreliable. During data acquisition, there are many sources of uncertainties and errors, especially patient motion, which result in the same image position being representative of many different anatomical locations across time. Apart from motion, there are also other inherent uncertainties and noise associated with the measurement of DCE-MRI parameters, which contribute to the model-fitting error observed when trying to apply pharmacokinetic (PK) models to the data. In this thesis, a probabilistic, model-based registration and parameter estimation (MoRPE) framework for motion correction and PK-parameter estimation in DCE-MRI sequences is presented. The MoRPE framework is first compared with conventional motion correction methods on simulated data, and then applied to data from a clinical trial involving twenty colorectal cancer patients. On clinical data, the ability of MoRPE to discriminate between responders and non-responders to combined chemoand radiotherapy is tested, and found to be superior to other methods. The effect of incorporating different arterial input functions within MoRPE is also assessed. Following this, a quantitative analysis of the uncertainties associated with the different PK parameters is performed using a variational Bayes mathematical framework. This analysis provides a quantitative estimate of the extent to which motion correction affects the uncertainties associated with different parameters. Finally, the importance of estimating spatial heterogeneity of PK parameters within tumours is assessed. The efficacy of different measures of spatial heterogeneity, in predicting response to therapy based on the pre-therapy scan alone are compared, and the prognostic value of a new derived PK parameter the 'acceleration constant' is investigated. The integration of uncertainty estimates of different DCE-MRI parameters into the calculation of their heterogeneity measures is also shown to improve the prediction of response to therapy.

The problems under consideration center around the interpretation of binocular stereo disparity. In particular, the goal is to establish a set of mappings from stereo disparity to corresponding three-dimensional scene geometry. An analysis has been developed that shows how disparity information can be interpreted in terms of three-dimensional scene properties, such as surface depth, discontinuities, and orientation. These theoretical developments have been embodied in a set of computer algorithms for the recovery of scene geometry from input stereo disparity. The results of applying these algorithms to several disparity maps are presented. Comparisons are made to the interpretation of stereo disparity by biological systems.

No description available.

Lung cancer is a disease affecting millions of people every year and poses a serious threat to global public health. Accurate lung cancer staging is crucial to choose an appropriate treatment protocol and to determine prognosis, this requires the acquisition of contrast-enhanced diagnostic CT (d-CT) that is usually followed by a PET/CT scan. Information from both d-CT and PET scan is used by the clinician in the staging process; however, these images are not intrinsically aligned because they are acquired on different days and on different scanners. Establishing anatomical correspondence, i.e., aligning the d-CT and the PET images is an inherently difficult task due to the absence of a direct relationship between the intensities of the images. The CT acquired during the PET/CT scan is used for attenuation correction (AC-CT) and is implicitly aligned with the PET image as they are acquired at the same time using a hybrid scanner. Patients are required to maintain shallow breathing for both scans. In contrast to that, the d-CT image is acquired after the injection of a contrast agent, and patients are required to maximally inhale, for better view of the lungs. Differences in the AC-CT and d-CT image volumes are thus due to differences in breathhold positions and image contrast. Nonetheless, both images are from the same modality. In this thesis, we present a new approach that aligns the d-CT with the PET image through an indirect registration process that uses the AC-CT. The deformation field obtained after the registration of the AC-CT to d-CT is used to align the PET image to the d-CT. Conventional image registration techniques deform the entire image using homogeneous regularization without taking into consideration the physical properties of the various anatomical structures. This homogeneous regularization may lead to physiologically and physically implausible deformations. To register the d-CT and AC-CT images, we developed a 3D registration framework based on a fluid transformation model including three physically motivated properties: (i) sliding motion of the lungs against the pleura; (ii) preservation of rigid structures; and (iii) preservation of topology. The sliding motion is modeled using a direction dependent regularization that decouples the tangential and the normal components of the external force term. The rigid shape of the bones is preserved using a spatially varying filter for the deformations. Finally, the topology is maintained using the concept of log-unbiased deformations. To solve the multi-modal registration problem due to the contrast injected for the d-CT, but lack thereof in the AC-CT, we use local cross correlation (LCC) as the similarity measure. To illustrate and validate the proposed registration framework, different intra-patient CT datasets are used, including the NCAT phantom, EMPIRE10 and POPI datasets. Results show that our proposed registration framework provides improved alignment and physically motivated deformations when compared to the classic elastic and fluid registration techniques. The final goal of our work was to demonstrate the clinical utility of our new approach that aligns d-CT and PET/AC-CT images for fusion. We apply our method to ten real patients. Our results show that the PET images have much improved alignment with the d-CT images using our proposed registration technique. Our method was successful in providing a good overlap of the lungs, improved alignment of the tumours and a lower target registration error for landmarks in comparison to the classic fluid registration. The main contribution of this thesis is the development of a comprehensive registration framework that integrates important physical properties into a state-of-the-art transformation model with application to lung imaging in cancer.

Object category detection, the task of determining if one or more instances of a category are present in an image with their corresponding locations, is one of the fundamental problems of computer vision. The task is very challenging because of the large variations in imaged object appearance, particularly due to the changes in viewpoint, illumination and intra-class variance. Although successful solutions exist for learning object category detectors, they require massive amounts of training data. Transfer learning builds upon previously acquired knowledge and thus reduces training requirements. The objective of this work is to develop and apply novel transfer learning techniques specific to the object category detection problem. This thesis proposes methods which not only address the challenges of performing transfer learning for object category detection such as finding relevant sources for transfer, handling aspect ratio mismatches and considering the geometric relations between the features; but also enable large scale object category detection by quickly learning from considerably fewer training samples and immediate evaluation of models on web scale data with the help of part-based indexing. Several novel transfer models are introduced such as: (a) rigid transfer for transferring knowledge between similar classes, (b) deformable transfer which tolerates small structural changes by deforming the source detector while performing the transfer, and (c) part level transfer particularly for the cases where full template transfer is not possible due to aspect ratio mismatches or not having adequately similar sources. Building upon the idea of using part-level transfer, instead of performing an exhaustive sliding window search, part-based indexing is proposed for efficient evaluation of templates enabling us to obtain immediate detection results in large scale image collections. Furthermore, easier and more robust optimization methods are developed with the help of feature maps defined between proposed transfer learning formulations and the “classical” SVM formulation.