Karin Engel

Closed-Form Hierarchical Finite Element Models for Part-Based Object Detection

In this work we address part-based object detection under variability of part shapes and spatial relations. Our approach bases on the hierarchical finite element modeling concept of Engel and Tönnies [ET09a, ET09b]. They model object parts by elastic materials, which adapt to image structures via image-derived forces. Spatial part relations are realized through additional layers of elastic material forming an elastic hierarchy. We present a closed-form solution to this concept, reformulating the hierarchical optimization problem into the optimization of a non-hierarchical finite element model. This allows us to apply standard finite element techniques to hierarchical problems and to provide an efficient framework for part-based object detection. We demonstrate our approach at the example of lumbar column detection in magnetic resonance imaging on a data set of 49 subjects. Given a rough model initialization, our approach solved the detection problem reliably in 45 out of 49 cases, showing computation times of only a few seconds per subject.

Efficient image segmentation using pairwise pixel similarities

Image segmentation based on pairwise pixel similarities has been a very active field of research in recent years. The drawbacks common to these segmentation methods are the enormous space and processor requirements. The contribution of this paper is a general purpose two-stage preprocessing method that substantially reduces the involved costs. Initially, an oversegmentation into small coherent image patches - or superpixels - is obtained through an iterative process guided by pixel similarities. A suitable pairwise superpixel similarity measure is then defined which may be plugged into an arbitrary segmentation method based on pairwise pixel similarities. To illustrate our ideas we integrated the algorithm into a spectral graph-partitioning method using the Normalized Cut criterion. Our experiments show that the time and memory requirements are reduced drastically (> 99%), while segmentations of adequate quality are obtained.

Hierarchical vibrations for part-based recognition of complex objects

We propose a technique for the recognition and segmentation of complex shapes in 2D images using a hierarchy of finite element vibration modes in an evolutionary shape search. The different levels of the shape hierarchy can influence each other, which can be exploited in top-down part-based image analysis. Our method overcomes drawbacks of existing structural approaches, which cannot uniformly encode shape variation and co-variation, or rely on training. We present results demonstrating that by utilizing a quality-of-fit function the model explicitly recognizes missing parts of a complex shape, thus allowing for categorization between shape classes.

Part-based localisation and segmentation of landmark-related auditory cortical regions

We recently presented a method for the delineation of cortical regions of interest that relies on the finite element decomposition of shape [21]. Our current work strengthens and extends the proposed technique with the following contributions: First, we provide a detailed discussion of the computational challenges related to applying the hierarchical shape modelling and energy minimisation approach to the representation and segmentation of specific areas in cortical surfaces. Second, we analyse the underlying heuristics in order to elucidate the representational power and accuracy of the a priori constrained, partial model of the auditory cortex anatomy, and improve the cortical landmark localisation. We show experimentally that a valid parametric prior can be built from expert prior knowledge in a straightforward manner. By employing the advantages of the hierarchical shape decomposition, the model can be substantially improved on the basis of training sets, which are much smaller compared with state-of-the-art methods.

Hierarchical Vibrations: A Structural Decomposition Approach for Image Analysis

We present results demonstrating that using a hierarchy of finite element vibration modes in an evolutionary deformable shape search provides a new interesting approach for the localization and segmentation of specific objects in 2D images. The design and coupling of the different levels of the shape hierarchy results in a multi---resolution shape space, which can be exploited in top---down part---based shape matching. The proposed strategy allows for segmenting complex objects from images, classification, as well as localization of the desired object under occlusions. It avoids misregistration by resolving several drawbacks inherent to standard shape---based approaches, which either cannot adequately represent non---linear variations, or rely on exhaustive prior training.

Automatic Segmentation of the Cortical Grey and White Matter in MRI Using a Region-Growing Approach Based on Anatomical Knowledge

We propose an automatic procedure for the correct segmentation of grey and white matter in MR data sets of the human brain. Our method exploits general anatomical knowledge for the initial segmentation and for the subsequent refinement of the estimation of the cortical grey matter. Our results are comparable to manual segmentations.

Parcellation of the Auditory Cortex into Landmark---Related Regions of Interest

We propose a method for the automated delineation of cortical regions of interest as a basis for the anatomo---functional parcellation of the human auditory cortex using neuroimaging. Our algorithm uses the properties of the cortical surface, and employs a recent hierarchical part---based pattern recognition strategy for a semantically correct labelling of the temporal lobe. The anatomical landmarks are finally combined to obtain an accurate separation and parametrisation of two auditory cortical regions. Experimental results show the good performance of the approach that was automated using simplified atlas information.

Model-Based Segmentation of Cortical Regions of Interest for Multi-subject Analysis of fMRI Data

The high inter-subject variability of human neuroanatomy complicates the analysis of functional imaging data across subjects. We propose a method for the correct segmentation of cortical regions of interest based on the cortical surface. First results on the segmentation of Heschl’s gyrus indicate the capability of our approach for correct comparison of functional activations in relation to individual cortical patterns.

Stable Structural Deformations

Recently, we introduced a hierarchical finite element model in the context of structural image segmentation. Such model deforms from its equilibrium shape into similar shapes under the influence of both, image---based forces and structural forces, which serve the propagation of deformations across the hierarchy levels. Such forces are very likely to result in large (rotational) deformations, which yield under the linear elasticity model artefacts and thus poor segmentation results. In this paper, we provide results indicating that different implementations of the stiffness warping method can be successfully combined to simulate dependent rotational deformations correctly, and in an efficient manner.

Fuzzy Multiscale Region Growing for Segmentation of MR Images of the Human Brain

We propose an automatic region growing technique for the segmentation of the cerebral cortex and white matter in MRI data. Our method exploits general anatomical knowledge and uses an iterative multi resolution scheme for the estimation of intensity distributions to compensate for artifacts within the data. We present a comparison to segmentation results created by the neuroimaging software Brainvoyager QX and show advantages of our approach based on a qualitative and quantitative evaluation.