Citation Context

...I data acquired per patient leads to several Proceedings IWBBIO 2014. Granada 7-9 April, 2014 798 2 complexities associated with the annotation of images and extraction of markers to differentiate the various subtle and rare forms of cardiomyopathies. These complexities necessitate the development of efficient workflows to acquire, manage and transform this data into actionable knowledge for patient care. Past work in this area has typically been in the image processing domain, wherein the effort has gone into imaging bio-marker extraction, segmentation of global and local regions of interest [1], extraction of quantitative metrics such as volume estimation, morphological, functional or flow-based features [2, 3] and finally automated tools for multi-modal image registration. To our knowledge, the clinical workflows associated with the CMRI data acquisition and processing have not been studied from a machine learning perspective to identify areas of inefficiencies wherein intelligent computational tools could be developed to aid cardiac radiologists in the assessment of cardiomyopathies using a multitude of imaging biomarkers. In this paper, we develop and test a novel framework calle...

Citation Context

...urface rendering by implementing a marching cubes algorithm [9]. Figure 4. 3D reconstructions of the heart structures correspondent to two patients: yellow-RV, red-LV with coronaries, green- myocardium, white-scar and violet- fat. On the left, the patient has little fat layer, on the right side instead, there is a thick fat layer. 4. Discussion Due to the technological development of MDCT in the last years, manual segmentation of such images is becoming an even less efficient option in the clinical practice. Today, model-based segmentation methods are widely used to segment cardiac structures [10, 11, 12, 13, 14]. However, they present some limitations on the LV and myocardium. In the former, papillary muscles were considered as LV cavity and in the latter, the substantial variation of wall thickness in such pathological subjects reduces their good performance. Therefore, it is usually still necessary to manually segment (or/and correct) the cardiac structures to plan adequately the ablation procedure. Additionally, to the best of our knowledge, there is no automatically segmentation method for epicardial fat and scar for ANGIO and LE images respectively, as proposed in this research. In this work we ...

Citation Context

...lities (such as ultrasound and magnetic resonancesimaging), cardiac computed tomography (CT) can provide detailed anatomical informationsabout the heart chambers, great vessels, and coronary arteries =-=[4,5]-=-. Actually, CT is oftenspreferred by diagnosticians since it provides more accurate anatomical information about thesvisualized structures, thanks to its higher signal-to-noise ratio and better spatia...