Space and Time Shape Constrained

The aim of this work is to automatically extract quantitative parameters from time sequences of 3D images (4D images) suited to heart pathology diagnosis. In this paper, we propose a framework for the reconstruction of the left ventricle motion from 4D images based on 4D deformable surface models. These 4D models are represented as a time sequence of 3D meshes whose deformation are correlated during the cardiac cycle. Both temporal and spatial constraints based on prior knowledge of heart shape and motion are combined to improve the segmentation accuracy. In contrast to many earlier approaches, our framework includes the notion of trajectory constraint. We have demonstrated the ability of this segmentation tool to deal with noisy or low contrast images on 4D MR, SPECT, and US images. 1 Context Recently, the improvement of medical image acquisition technology has allowed the production of time sequences of 3D medical images (4D images) for several image modalities (CT, MRI, US, SPECT). Tagged MRI is the gold standard of heart motion analysis since it is the only modality permitting the extraction of the motion of physical points located in the myocardium 18]. However, other modalities may be used for meaningful parameters extraction at a lower cost. In particular, the fast development of 3D US imaging is very promising due to its accessibility and low cost 17]. The main target for these new ultra-fast image acquisition devices is to capture and analyze the heart motion through the extraction of quantitative parameters such as volume, walls thickness, ejection fraction and motion amplitude. In order to estimate these parameters, it is necessary to reconstruct the Left Ventricle (LV) motion during a cardiac cycle. Tracking the LV motion in 2D or 3D image sequences has led to several research eeorts 10, 9, 2]. Tracking 12, 16] and motion analysis 4, 7] based on deformable models in 4D images take into account time continuity and periodicity to improve their robustness. In this paper, we propose to track the LV based on 4D deformable models. Our concept of 4D deformable surface models combines spatial and temporal constraints which diiers from most previous approaches 12, 16, 4] that decouple them. Furthermore, in contrast to the strategy presented in 7], the motion estimation is not parameterized by a global time-space transformation. It leads to