Abstract: A through description of the left ventricle functionality requires combining complementary regional scores. A main limitation is the lack of multiparametric normality models oriented to the assessment of regional wall motion abnormalities (RWMA). This paper covers two main topics involved in RWMA assessment. We propose a general framework allowing the fusion and comparison across subjects of different regional scores. Our framework is used to explore which combination of regional scores (including 2-D motion and strains) is better suited for RWMA detection. Our statistical analysis indicates that for a proper (within interobserver variability) identification of RWMA, models should consider motion and extreme strains.

Abstract: In this work, we present a fully automated method for tissue deformation estimation in tagged magnetic resonance images (TMRI). Gabor filter banks, tuned independently for each left ventricle level, provide optimally filtered complex images which phase remains constant along the cardiac cycle. This fact can be thought as the brightness constancy condition required by classical optical flow (OF) methods. Pairs of these filtered sequences, together with a variational formulation are used in a second step to obtain dense continuous deformation maps that we call Harmonic Phase Flow. This method has been used to determine reference values of ventricular torsion (VT) in a set of 8 healthy volunteers. The results encourage the use of VT as a useful parameter for ventricular function assessment in clinical routine.

Abstract: Medical imaging is very useful in the assessment and treatment of many diseases. To deal with the great amount of data provided by imaging scanners and extract quantitative information that physicians can interpret, many analysis algorithms have been developed. Any process of analysis always consists of a first step of segmenting some particular structure. In medical imaging, structures are not always well defined and suffer from noise artifacts thus, ordinary segmentation methods are not well suited. The ones that seem to give better results are those based on deformable models. Nevertheless, despite their capability of mixing image features together with smoothness constraints that may compensate for image irregularities, these are naturally local methods, i. e., each node of the active contour evolve taking into account information about its neighbors and some other weak constraints about flexibility and smoothness, but not about the global shape that they should find. Due to the fact that structures to be segmented are the same for all cases but with some inter and intra-patient variation, the incorporation of a priori knowledge about shape in the segmentation method will provide robustness to it. Active Shape Models is an algorithm based on the creation of a shape model called Point Distribution Model. It performs a segmentation using only shapes similar than those previously learned from a training set that capture most of the variation presented by the structure. This algorithm works by updating shape nodes along a normal segment which often can be too restrictive. For this reason we propose a generalization of this algorithm that we call Generalized Active Shape Models and fully integrates the a priori knowledge given by the Point Distribution Model with deformable models or any other appropriate segmentation method. Two different applications to cardiac imaging of this generalized method are developed and promising results are shown.

Abstract: Crystal unbending, the process that aims to recover a perfect crystal from experimental data, is one of the more important steps in electron crystallography image processing. The unbending process involves three steps: estimation of the unit cell displacements from their ideal positions, extension of the deformation field to the whole image and transformation of the image in order to recover an ideal crystal. In this work, we present a systematic analysis of the second step oriented to address two issues. First, whether the unit cells remain undistorted and only the distance between them should be changed (rigid case) or should be modified with the same deformation suffered by the whole crystal (elastic case). Second, the performance of different extension algorithms (interpolation versus approximation) is explored. Our experiments show that there is no difference between elastic and rigid cases or among the extension algorithms. This implies that the deformation fields are constant over large areas. Furthermore, our results indicate that the main source of error is the transformation of the crystal image.