Advanced Kernel Methods For Remote Sensing Image Classification

Résumé Suite aux recentes avancées technologiques, les archives d'images digitales ont connu une croissance qualitative et quantitative sans précédent. Malgré les énormes possibilités qu'elles offrent, ces avancées posent de nouvelles questions quant au traitement des masses de données saisies. Cette question est à la base de cette Thèse: les problèmes de traitement d'information digitale à très haute résolution spatiale et/ou spectrale y sont considérés en recourant à des approches d'apprentissage statistique, les méthodes à noyau. Cette Thèse étudie des problèmes de classification d'images, c'est à dire de catégorisation de pixels en un nombre réduit de classes refletant les propriétés spectrales et contextuelles des objets qu'elles représentent. L'accent est mis sur l'efficience des algorithmes, ainsi que sur leur simplicité, de manière à augmenter leur potentiel d'implementation pour les utilisateurs. De plus, le défi de cette Thèse est de rester proche des problèmes concrets des utilisateurs d'images satellite sans pour autant perdre de vue l'intéret des méthodes proposées pour le milieu du machine learning dont elles sont issues. En ce sens, ce travail joue la carte de la transdisciplinarité en maintenant un lien fort entre les deux sciences dans tous les développements proposés. Quatre modèles sont proposés: le premier répond au problème de la haute dimensionalité et de la redondance des données par un modèle optimisant les performances en classification en s'adaptant aux particularités de l'image. Ceci est rendu possible par un système de ranking des variables (les bandes) qui est optimisé en même temps que le modèle de base: ce faisant, seules les variables importantes pour résoudre le problème sont utilisées par le classifieur. Le manque d'information étiquétée et l'incertitude quant à sa pertinence pour le problème sont à la source des deux modèles suivants, basés respectivement sur l'apprentissage actif et les méthodes semi-supervisées: le premier permet d'améliorer la qualité d'un ensemble d'entraînement par interaction directe entre l'utilisateur et la machine, alors que le deuxième utilise les pixels non étiquetés pour améliorer la description des données disponibles et la robustesse du modèle. Enfin, le dernier modèle proposé considère la question plus théorique de la structure entre les outputs: l'intègration de cette source d'information, jusqu'à présent jamais considérée en télédétection, ouvre des nouveaux défis de recherche. Advanced kernel methods for remote sensing image classification Devis Tuia Institut de Géomatique et d'Analyse du Risque September 2009 Abstract The technical developments in recent years have brought the quantity and quality of digital information to an unprecedented level, as enormous archives of satellite images are available to the users. However, even if these advances open more and more possibilities in the use of digital imagery, they also rise several problems of storage and treatment. The latter is considered in this Thesis: the processing of very high spatial and spectral resolution images is treated with approaches based on data-driven algorithms relying on kernel methods. In particular, the problem of image classification, i.e. the categorization of the image's pixels into a reduced number of classes reflecting spectral and contextual properties, is studied through the different models presented. The accent is put on algorithmic efficiency and the simplicity of the approaches proposed, to avoid too complex models that would not be used by users. The major challenge of the Thesis is to remain close to concrete remote sensing problems, without losing the methodological interest from the machine learning viewpoint: in this sense, this work aims at building a bridge between the machine learning and remote sensing communities and all the models proposed have been developed keeping in mind the need for such a synergy. Four models are proposed: first, an adaptive model learning the relevant image features has been proposed to solve the problem of high dimensionality and collinearity of the image features. This model provides automatically an accurate classifier and a ranking of the relevance of the single features. The scarcity and unreliability of labeled. information were the common root of the second and third models proposed: when confronted to such problems, the user can either construct the labeled set iteratively by direct interaction with the machine or use the unlabeled data to increase robustness and quality of the description of data. Both solutions have been explored resulting into two methodological contributions, based respectively on active learning and semisupervised learning. Finally, the more theoretical issue of structured outputs has been considered in the last model, which, by integrating outputs similarity into a model, opens new challenges and opportunities for remote sensing image processing.

ZHARP: three-dimensional motion tracking from a single image plane.

Three-dimensional imaging and quantification of myocardial function are essential steps in the evaluation of cardiac disease. We propose a tagged magnetic resonance imaging methodology called zHARP that encodes and automatically tracks myocardial displacement in three dimensions. Unlike other motion encoding techniques, zHARP encodes both in-plane and through-plane motion in a single image plane without affecting the acquisition speed. Postprocessing unravels this encoding in order to directly track the 3-D displacement of every point within the image plane throughout an entire image sequence. Experimental results include a phantom validation experiment, which compares zHARP to phase contrast imaging, and an in vivo study of a normal human volunteer. Results demonstrate that the simultaneous extraction of in-plane and through-plane displacements from tagged images is feasible.

Performance of a new gadolinium-based intravascular contrast agent in free-breathing inversion-recovery 3D coronary MRA.

In three-dimensional (3D) coronary magnetic resonance angiography (MRA), the in-flow contrast between the coronary blood and the surrounding myocardium is attenuated as compared to thin-slab two-dimensional (2D) techniques. The application of a gadolinium (Gd)-based intravascular contrast agent may provide an additional source of signal and contrast by reducing T(1blood) and supporting the visualization of more distal or branching segments of the coronary arterial tree. In six healthy adults, the left coronary artery (LCA) system was imaged pre- and postcontrast with a 0.075-mmol/kg bodyweight dose of the intravascular contrast agent B-22956. For imaging, an optimized free-breathing, navigator-gated and -corrected 3D inversion recovery (IR) sequence was used. For comparison, state-of-the-art baseline 3D coronary MRA with T(2) preparation for non-exogenous contrast enhancement was acquired. The combination of IR 3D coronary MRA, sophisticated navigator technology, and B-22956 allowed for an extensive visualization of the LCA system. Postcontrast, a significant increase in both the signal-to-noise ratio (SNR; 46%, P < 0.05) and contrast-to-noise ratio (CNR; 160%, P < 0.01) was observed, while vessel sharpness of the left anterior descending (LAD) artery and the left coronary circumflex (LCX) were improved by 20% (P < 0.05) and 18% (P < 0.05), respectively.

Irritability in pre-clinical Huntington's disease.

Irritability, together with depression and anxiety, form three salient clinical features of pre-symptomatic Huntington's disease (HD). To date, the understanding of irritability in HD suffers from a paucity of experimental data and is largely based on questionnaires or clinical anecdotes. Factor analysis suggests that irritability is related to impulsivity and aggression and is likely to engage the same neuronal circuits as these behaviours, including areas such as medial orbitofrontal cortex (OFC) and amygdala. 16 pre-symptomatic gene carriers (PSCs) and 15 of their companions were asked to indicate the larger of two squares consecutively shown on a screen while undergoing functional magnetic resonance imaging (fMRI). Despite correct identification of the larger square, participants were often told that they or their partner had given the wrong answer. Size differences were subtle to make negative feedback credible but detectable. Although task performance, baseline irritability, and reported task-induced irritation were the same for both groups, fMRI revealed distinct neuronal processing in those who will later develop HD. In controls but not PSCs, task-induced irritation correlated positively with amygdala activation and negatively with OFC activation. Repetitive negative feedback induced greater amygdala activations in controls than PSCs. In addition, the inverse functional coupling between amygdala and OFC was significantly weaker in PSCs compared to controls. Our results argue that normal emotion processing circuits are disrupted in PSCs via attenuated modulation of emotional status by external or internal indicators. At later stages, this dysfunction may increase the risk for developing recognised, HD-associated, psychiatric symptoms such as irritability.

Objective assessment of image quality in conventional and digital mammography taking into account dynamic range.

The goal of this work is to develop a method to objectively compare the performance of a digital and a screen-film mammography system in terms of image quality. The method takes into account the dynamic range of the image detector, the detection of high and low contrast structures, the visualisation of the images and the observer response. A test object, designed to represent a compressed breast, was constructed from various tissue equivalent materials ranging from purely adipose to purely glandular composition. Different areas within the test object permitted the evaluation of low and high contrast detection, spatial resolution and image noise. All the images (digital and conventional) were captured using a CCD camera to include the visualisation process in the image quality assessment. A mathematical model observer (non-prewhitening matched filter), that calculates the detectability of high and low contrast structures using spatial resolution, noise and contrast, was used to compare the two technologies. Our results show that for a given patient dose, the detection of high and low contrast structures is significantly better for the digital system than for the conventional screen-film system studied. The method of using a test object with a large tissue composition range combined with a camera to compare conventional and digital imaging modalities can be applied to other radiological imaging techniques. In particular it could be used to optimise the process of radiographic reading of soft copy images.

Fast Geodesic Active Fields for Image Registration Based on Splitting and Augmented Lagrangian Approaches.

In this paper, we present an efficient numerical scheme for the recently introduced geodesic active fields (GAF) framework for geometric image registration. This framework considers the registration task as a weighted minimal surface problem. Hence, the data-term and the regularization-term are combined through multiplication in a single, parametrization invariant and geometric cost functional. The multiplicative coupling provides an intrinsic, spatially varying and data-dependent tuning of the regularization strength, and the parametrization invariance allows working with images of nonflat geometry, generally defined on any smoothly parametrizable manifold. The resulting energy-minimizing flow, however, has poor numerical properties. Here, we provide an efficient numerical scheme that uses a splitting approach; data and regularity terms are optimized over two distinct deformation fields that are constrained to be equal via an augmented Lagrangian approach. Our approach is more flexible than standard Gaussian regularization, since one can interpolate freely between isotropic Gaussian and anisotropic TV-like smoothing. In this paper, we compare the geodesic active fields method with the popular Demons method and three more recent state-of-the-art algorithms: NL-optical flow, MRF image registration, and landmark-enhanced large displacement optical flow. Thus, we can show the advantages of the proposed FastGAF method. It compares favorably against Demons, both in terms of registration speed and quality. Over the range of example applications, it also consistently produces results not far from more dedicated state-of-the-art methods, illustrating the flexibility of the proposed framework.

Influence of magnetic field strength and image registration strategy on voxel-based morphometry in a study of Alzheimer's disease.

Multi-centre data repositories like the Alzheimer's Disease Neuroimaging Initiative (ADNI) offer a unique research platform, but pose questions concerning comparability of results when using a range of imaging protocols and data processing algorithms. The variability is mainly due to the non-quantitative character of the widely used structural T1-weighted magnetic resonance (MR) images. Although the stability of the main effect of Alzheimer's disease (AD) on brain structure across platforms and field strength has been addressed in previous studies using multi-site MR images, there are only sparse empirically-based recommendations for processing and analysis of pooled multi-centre structural MR data acquired at different magnetic field strengths (MFS). Aiming to minimise potential systematic bias when using ADNI data we investigate the specific contributions of spatial registration strategies and the impact of MFS on voxel-based morphometry in AD. We perform a whole-brain analysis within the framework of Statistical Parametric Mapping, testing for main effects of various diffeomorphic spatial registration strategies, of MFS and their interaction with disease status. Beyond the confirmation of medial temporal lobe volume loss in AD, we detect a significant impact of spatial registration strategy on estimation of AD related atrophy. Additionally, we report a significant effect of MFS on the assessment of brain anatomy (i) in the cerebellum, (ii) the precentral gyrus and (iii) the thalamus bilaterally, showing no interaction with the disease status. We provide empirical evidence in support of pooling data in multi-centre VBM studies irrespective of disease status or MFS.

Model-based Segmentation and Image Fusion of 3D Computed Tomography and 3D Ultrasound of the Eye for Radiotherapy Planning

For radiotherapy treatment planning of retinoblastoma inchildhood, Computed Tomography (CT) represents thestandard method for tumor volume delineation, despitesome inherent limitations. CT scan is very useful inproviding information on physical density for dosecalculation and morphological volumetric information butpresents a low sensitivity in assessing the tumorviability. On the other hand, 3D ultrasound (US) allows ahigh accurate definition of the tumor volume thanks toits high spatial resolution but it is not currentlyintegrated in the treatment planning but used only fordiagnosis and follow-up. Our ultimate goal is anautomatic segmentation of gross tumor volume (GTV) in the3D US, the segmentation of the organs at risk (OAR) inthe CT and the registration of both. In this paper, wepresent some preliminary results in this direction. Wepresent 3D active contour-based segmentation of the eyeball and the lens in CT images; the presented approachincorporates the prior knowledge of the anatomy by usinga 3D geometrical eye model. The automated segmentationresults are validated by comparing with manualsegmentations. Then, for the fusion of 3D CT and USimages, we present two approaches: (i) landmark-basedtransformation, and (ii) object-based transformation thatmakes use of eye ball contour information on CT and USimages.

Inexperienced sonographers can successfully visualize and assess a three-dimensional image of the fetal face using a standardized ultrasound protocol.

Introduction: A standardized three-dimensional ultrasonographic (3DUS) protocol is described that allows fetal face reconstruction. Ability to identify cleft lip with 3DUS using this protocol was assessed by operators with minimal 3DUS experience. Material and Methods: 260 stored volumes of fetal face were analyzed using a standardized protocol by operators with different levels of competence in 3DUS. The outcomes studied were: (1) the performance of post-processing 3D face volumes for the detection of facial clefts; (2) the ability of a resident with minimal 3DUS experience to reconstruct the acquired facial volumes, and (3) the time needed to reconstruct each plane to allow proper diagnosis of a cleft. Results: The three orthogonal planes of the fetal face (axial, sagittal and coronal) were adequately reconstructed with similar performance when acquired by a maternal-fetal medicine specialist or by residents with minimal experience (72 vs. 76%, p = 0.629). The learning curve for manipulation of 3DUS volumes of the fetal face corresponds to 30 cases and is independent of the operator's level of experience. Discussion: The learning curve for the standardized protocol we describe is short, even for inexperienced sonographers. This technique might decrease the length of anatomy ultrasounds and improve the ability to visualize fetal face anomalies.

Functional compensation of motor function in pre-symptomatic Huntington's disease.

Involuntary choreiform movements are a clinical hallmark of Huntington's disease. Studies in clinically affected patients suggest a shift of motor activations to parietal cortices in response to progressive neurodegeneration. Here, we studied pre-symptomatic gene carriers to examine the compensatory mechanisms that underlie the phenomenon of retained motor function in the presence of degenerative change. Fifteen pre-symptomatic gene carriers and 12 matched controls performed button presses paced by a metronome at either 0.5 or 2 Hz with four fingers of the right hand whilst being scanned with functional magnetic resonance imaging. Subjects pressed buttons either in the order of a previously learnt 10-item finger sequence, from left to right, or kept still. Error rates ranged from 2% to 7% in the pre-symptomatic gene carriers and from 0.5% to 4% in controls, depending on the condition. No significant difference in task performance was found between groups for any of the conditions. Activations in the supplementary motor area (SMA) and superior parietal lobe differed with gene status. Compared with healthy controls, gene carriers showed greater activations of left caudal SMA with all movement conditions. Activations correlated with increasing speed of movement were greater the closer the gene carriers were to estimated clinical diagnosis, defined by the onset of unequivocal motor signs. Activations associated with increased movement complexity (i.e. with the pre-learnt 10-item sequence) decreased in the rostral SMA with nearing diagnostic onset. The left superior parietal lobe showed reduced activation with increased movement complexity in gene carriers compared with controls, and in the right superior parietal lobe showed greater activations with all but the most demanding movements. We identified a complex pattern of motor compensation in pre-symptomatic gene carriers. The results show that preclinical compensation goes beyond a simple shift of activity from premotor to parietal regions involving multiple compensatory mechanisms in executive and cognitive motor areas. Critically, the pattern of motor compensation is flexible depending on the actual task demands on motor control.

Flow-targeted inversion-prepared b-TFE coronary MR angiography: initial results in patients.

PURPOSE: Visualization of coronary blood flow in the right and left coronary system in volunteers and patients by means of a modified inversion-prepared bright-blood coronary magnetic resonance angiography (cMRA) sequence. MATERIALS AND METHODS: cMRA was performed in 14 healthy volunteers and 19 patients on a 1.5 Tesla MR system using a free-breathing 3D balanced turbo field echo (b-TFE) sequence with radial k-space sampling. For magnetization preparation a slab selective and a 2D selective inversion pulse were used for the right and left coronary system, respectively. cMRA images were evaluated in terms of clinically relevant stenoses (< 50 %) and compared to conventional catheter angiography. Signal was measured in the coronary arteries (coro), the aorta (ao) and in the epicardial fat (fat) to determine SNR and CNR. In addition, maximal visible vessel length, and vessel border definition were analyzed. RESULTS: The use of a selective inversion pre-pulse allowed direct visualization of the coronary blood flow in the right and left coronary system. The measured SNR and CNR, vessel length, and vessel sharpness in volunteers (SNR coro: 28.3 +/- 5.0; SNR ao: 37.6 +/- 8.4; CNR coro-fat: 25.3 +/- 4.5; LAD: 128.0 cm +/- 8.8; RCA: 74.6 cm +/- 12.4; Sharpness: 66.6 % +/- 4.8) were slightly increased compared to those in patients (SNR coro: 24.1 +/- 3.8; SNR ao: 33.8 +/- 11.4; CNR coro-fat: 19.9 +/- 3.3; LAD: 112.5 cm +/- 13.8; RCA: 69.6 cm +/- 16.6; Sharpness: 58.9 % +/- 7.9; n.s.). In the patient study the assessment of 42 coronary segments lead to correct identification of 10 clinically relevant stenoses. CONCLUSION: The modification of a previously published inversion-prepared cMRA sequence allowed direct visualization of the coronary blood flow in the right as well as in the left coronary system. In addition, this sequence proved to be highly sensitive regarding the assessment of clinically relevant stenotic lesions.

Image quality assessment in digital mammography: part II. NPWE as a validated alternative for contrast detail analysis.

Assessment of image quality for digital x-ray mammography systems used in European screening programs relies mainly on contrast-detail CDMAM phantom scoring and requires the acquisition and analysis of many images in order to reduce variability in threshold detectability. Part II of this study proposes an alternative method based on the detectability index (d') calculated for a non-prewhitened model observer with an eye filter (NPWE). The detectability index was calculated from the normalized noise power spectrum and image contrast, both measured from an image of a 5 cm poly(methyl methacrylate) phantom containing a 0.2 mm thick aluminium square, and the pre-sampling modulation transfer function. This was performed as a function of air kerma at the detector for 11 different digital mammography systems. These calculated d' values were compared against threshold gold thickness (T) results measured with the CDMAM test object and against derived theoretical relationships. A simple relationship was found between T and d', as a function of detector air kerma; a linear relationship was found between d' and contrast-to-noise ratio. The values of threshold thickness used to specify acceptable performance in the European Guidelines for 0.10 and 0.25 mm diameter discs were equivalent to threshold calculated detectability indices of 1.05 and 6.30, respectively. The NPWE method is a validated alternative to CDMAM scoring for use in the image quality specification, quality control and optimization of digital x-ray systems for screening mammography.

Free-breathing whole-heart coronary MRA with 3D radial SSFP and self-navigated image reconstruction.

Respiratory motion is a major source of artifacts in cardiac magnetic resonance imaging (MRI). Free-breathing techniques with pencil-beam navigators efficiently suppress respiratory motion and minimize the need for patient cooperation. However, the correlation between the measured navigator position and the actual position of the heart may be adversely affected by hysteretic effects, navigator position, and temporal delays between the navigators and the image acquisition. In addition, irregular breathing patterns during navigator-gated scanning may result in low scan efficiency and prolonged scan time. The purpose of this study was to develop and implement a self-navigated, free-breathing, whole-heart 3D coronary MRI technique that would overcome these shortcomings and improve the ease-of-use of coronary MRI. A signal synchronous with respiration was extracted directly from the echoes acquired for imaging, and the motion information was used for retrospective, rigid-body, through-plane motion correction. The images obtained from the self-navigated reconstruction were compared with the results from conventional, prospective, pencil-beam navigator tracking. Image quality was improved in phantom studies using self-navigation, while equivalent results were obtained with both techniques in preliminary in vivo studies.