Skip to Main content Skip to Navigation

Probabilistic atlas statistical estimation with multimodal datasets and its application to atlas based segmentation

Abstract : Computerized anatomical atlases play an important role in medical image analysis. While an atlas usually refers to a standard or mean image also called template, that presumably represents well a given population, it is not enough to characterize the observed population in detail. A template image should be learned jointly with the geometric variability of the shapes represented in the observations. These two quantities will in the sequel form the atlas of the corresponding population. The geometric variability is modelled as deformations of the template image so that it fits the observations. In the first part of the work, we provide a detailed analysis of a new generative statistical model based on dense deformable templates that represents several tissue types observed in medical images. Our atlas contains both an estimation of probability maps of each tissue (called class) and the deformation metric. We use a stochastic algorithm for the estimation of the probabilistic atlas given a dataset. This atlas is then used for atlas-based segmentation method to segment the new images. Experiments are shown on brain T1 MRI datasets. Traditional analyses of Functional Magnetic Resonance Imaging use little anatomical information. The registration of the images to a template is based on the individual anatomy and ignores functional information; subsequently detected activations are not confined to gray matter. In the second part of the work, we propose a statistical model to estimate a probabilistic atlas from functional and T1 MRIs that summarizes both anatomical and functional information and the geometric variability of the population. Registration and Segmentation are performed jointly along the atlas estimation and the functional activity is constrained to the gray matter, increasing the accuracy of the atlas. Inferring protein abundances from peptide intensities is the key step in quantitative proteomics. The inference is necessarily more accurate when many peptides are taken into account for a given protein. Yet, the information brought by the peptides shared by different proteins is commonly discarded. In the third part of the work, we propose a statistical framework based on a hierarchical modeling to include that information. Our methodology, based on a simultaneous analysis of all the quantified peptides, handles the biological and technical errors as well as the peptide effect. In addition, we propose a practical implementation suitable for analyzing large datasets. Compared to a method based on the analysis of one protein at a time (that does not include shared peptides), our methodology proved to be far more reliable for estimating protein abundances and testing abundance changes.
Complete list of metadata

Cited literature [147 references]  Display  Hide  Download
Contributor : Hao Xu Connect in order to contact the contributor
Submitted on : Tuesday, April 8, 2014 - 10:49:56 AM
Last modification on : Wednesday, October 20, 2021 - 12:23:58 AM


  • HAL Id : pastel-00969176, version 1



Hao Xu. Probabilistic atlas statistical estimation with multimodal datasets and its application to atlas based segmentation. Statistics [math.ST]. Ecole Polytechnique X, 2014. English. ⟨pastel-00969176⟩



Record views


Files downloads