Pyramide (traitement d'image) | EPFL Graph Search

En traitement d'images, la pyramide est une représentation multi-résolution d'une image. Elle permet de modéliser l'image à différentes , depuis l'image initiale jusqu'à une image très grossière. La pyramide d'images est souvent utilisée car elle permet à l'algorithme de traitement d'image de travailler depuis les détails jusqu'au « grossier ». Cet outil est notamment utilisé à des fins de . Les principaux types de construction d'une pyramide d'images sont :

Gaussienne
Laplacienne
Irrégulière
Adaptative

Génération d'une pyramide Il existe deux principaux types de pyramides : passe-bas, et passe-bande. Une pyramide passe-bas est réalisée en lissant l'image avec un filtre approprié pour ensuite la sous-échantillonner avec l'image lissée. On utilise généralement un facteur 2 le long de chaque direction. La procédure est ensuite répétée sur l'image résultante. Ce cycle est répété plusieurs fois. Chaque cycle de cette procédure résulte en une image plus petite avec

Sparse Binary Features for Image Classification

Johann Paratte

In this work a new method for automatic image classification is proposed. It relies on a compact representation of images using sets of sparse binary features. This work first evaluates the Fast Retina Keypoint binary descriptor and proposes improvements based on an efficient descriptor representation. The efficient representation is created using dimensionality reduction techniques, entropy analysis and decorrelated sampling. In a second part, the problem of image classification is tackled. The traditional approach uses machine learning algorithms to create classifiers, and some works already propose to use a compact image representation using feature extraction as preprocessing. The second contribution of this work is to show that binary features, while being very compact and low dimensional (compared to traditional representation of images), still provide a very high discriminant power. This is shown using various learning algorithms and binary descriptors. These years a scheme has been widely used to perform object recognition on images, or equivalently image classification. It is based on the concept of Bag of Visual Words. More precisely, an image is described using an unordered set of visual words, that are generally represented by feature descriptions. The last contribution of this work is to use binary features with a simple Bag of Visual Words classifier. Tests of performance for the image classification are performed on a large database of images.

2013

Nonrigid medical image registration

Francisco Javier Sanchez Castro

Over the last two decades medical image processing has become a fundamental tool in medicine, and brain imaging is a paradigm thereof. The fast development of the imaging acquisition techniques allows medical staff to dispose of a variety of data coming from different imaging modalities. However, the analysis of such amount of data based on visual inspection is very difficult and highly time consuming. Often, the information contained in a medical image cannot be entirely seen by a human eye and vice-versa, computers do not have human common sense or the experience acquired by medical experts. Therefore, it is desirable to combine both medical experts and computers in an optimal trade-off to improve the outcome of the wide range of applications where medical imaging is nowadays applied. Among these applications we can mention diagnosis, surgical planning and intra-operative surgical navigation. Image registration is one of the most popular image processing techniques used in clinical environments. It consists of bringing two or more images into a single spatial representation, i.e. finding the optimal geometric transformation between the data. Image registration overcomes the difficulties encountered by clinicians to fuse images mentally since it enables to integrate different images into one representation. The applications of image registration in medicine are numerous and can be grouped in clinical, related to detection and diagnosis, and surgical. Some of these applications are: preprocedural planning and simulation, interventional radiology, diagnostic radiology, minimally invasive procedures, radiation therapy, intraoperative navigation, robot-assisted interventions, etc. During the decade of the 90s a large number of registration algorithms were proposed, linked to a variety of medical applications. Nowadays, rigid registration is considered to be a well-covered domain with efficient solutions for the most common applications. Current research mostly focuses on nonrigid registration methods. However, the complex framework of nonrigid registration makes it a complex problem where there is not a universal solution. This complex framework includes a number of components that interact: the dimensionality of the data involved, the feature space to base the registration on, the kind and domain of the transformation model, the modalities of the data and the subjects involved, the user interaction, the similarity metric to assess the alignment quantitatively, the search space where the optimal solution has to be found, and the search strategy including optimization of the objective function and implementation details. We can summarize the motivation of this thesis in the following sentence: Validation of nonrigid registration algorithms is a very difficult task and rather an application-dependent problem. Therefore, this dissertation is structured to face the problem of brain image registration in three closely correlated parts: algorithms, applications and validation. In the algorithmic context, we propose a control point based nonrigid registration algorithm which aims to improve the performance of classical point based algorithms. In the framework of the implementation of the algorithm we develop a control point detector based on information theory tools and a transformation model based on a type of radial basis functions specially suited for interpolation of scattered and nonuniform distributed data samples. The algorithm is evaluated for typical inter-subject monomodal and multimodal brain image registration. In the context of clinical applications we face a neurosurgical application: the subthalamic nucleus targeting for deep brain stimulation in Parkinson's disease surgery. The extensive study covers from state-of-the-art targeting procedures to automatic targeting methods through different registration algorithms. Moreover, we analyze the influence of surrounding anatomical features on the target location. We also study the feasibility of an atlas of landmarks aiming to improve the targeting accuracy and at the same time to reduce both the computational cost and the preprocessing needed compared to the other automatic targeting methods proposed. All the proposed methods and techniques are statistically cross-validated against state-of-the-art techniques and expert's targeting variability through an application-adapted validation scheme. The results show that automatic subthalamic nucleus localization is possible and as accurate as the methods currently used. We also provide valuable conclusions about the anatomical structures that must be used to infer more accurate targets. In the context of validation of brain image registration, we develop a general (non application-oriented) validation scheme that provides an independent and high dimensional means of assessing the quality of the alignment given by a registration algorithm. This procedure is applied to structural magnetic resonance images and is based on the information carried by the diffusion tensor images. It provides an error measure per voxel allowing an independent evaluation for different regions. Almost 800 registered images are analyzed to provide with strength the conclusions derived from the statistical analysis.

EPFL2008

Variational methods for texture segmentation

Nawal Houhou

In the last decades, image production has grown significantly. From digital photographs to the medical scans, including satellite images and video films, more and more data need to be processed. Consequently the number of applications based on digital images has increased, either for medicine, research for country planning or for entertainment business such as animation or video games. All these areas, although very different one to another, need the same image processing techniques. Among all these techniques, segmentation is probably one of the most studied because of its important role. Segmentation is the process of extracting meaningful objects from an image. This task, although easily achieved by the human visual system, is actually complex and still a true challenge for the image processing community despite several decades of research. The thesis work presented in this manuscript proposes solutions to the image segmentation problem in a well established mathematical framework, i.e. variational models. The image is defined in a continuous space and the segmentation problem is expressed through a functional or energy optimization. Depending on the object to be segmented, this energy definition can be difficult; in particular for objects with ambiguous borders or objects with textures. For the latter, the difficulty lies already in the definition of the term texture. The human eye can easily recognize a texture, but it is quite difficult to find words to define it, even more in mathematical terms. There is a deliberate vagueness in the definition of texture which explains the difficulty to conceptualize a model able to describe it. Often these textures can neither be described by homogeneous regions nor by sharp contours. This is why we are first interested in the extraction of texture features, that is to say, finding one representation that can discriminate a textured region from another. The first contribution of this thesis is the construction of a texture descriptor from the representation of the image similar to a surface in a volume. This descriptor belongs to the framework of non-supervised segmentation, since it will not require any user interaction. The second contribution is a solution for the segmentation problem based on active contour models and information theory tools. Third contribution is a semi-supervised segmentation model, i.e. where constraints provided by the user will be integrated in the segmentation framework. This processus is actually derived from the graph of image patches. This graph gives the connectivity measure between the different points of the image. The segmentation will be expressed by a graph partition and a variational model. This manuscript proposes to tackle the segmentation problem for textured images.