Space and Time Shape Constrained Deformable Surfaces for 4D Medical Image Segmentation

The heart motion study is vital for the understanding and the early diagnosis of cardiac pathologies that remain the primary cause of death in western countries to date. Heart motion analysis require the extraction of quantitative parameters from time sequences of 3D images (4D images) such as volume, sep-tum wall thickness, ejection fraction and motion amplitude. 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 are combined to introduce a prior knowledge of the heart motion and improve the segmenta-tion accuracy. When compared to previous approaches, our framework appears too be more eecient and more powerful since it can also include the notion of trajectory constraint. We have validated this segmentation tool and its ability to segment even noisy or low contrasted images on 4D MR, SPECT and ultrasound images. 1 Context Recently, the improvement of medical image acquisition technology has allowed the production of time sequences of 3D medical images (3D+T or 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 ultrasound imaging is very promising for performing cardiac motion analysis 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, septum wall thickness, ejection fraction and motion amplitude. However, in order to estimate these parameters, it is necessary to reconstruct the left ventricle motion during a cardiac cycle. Tracking the ventricle motion in 2D or 3D image sequences has been the motivation for much research work 10, 9, 2]. Tracking 12, 16] and motion analysis 4, 7] based on deformable models in 4D images take into account the time continuity and periodicity to improve their robustness.