A Novel 4D Reconstruction Method Of Organs And Tissues Acquired By High Frequency Scans

Current CAD (Computer Aided Diagnosis) systems supporting the present medical imaging scanners (e.g, Computed Tomography, (CT), Magnetic Resonance Imaging, (MRI)) have different limitations, including movement artifacts, and 4D reconstruction of moving organs (e.g, heart and lungs). The new generation of medical imaging scanners (e.g., 4D CT) allows us to acquire simultaneous multidirectional images (i.e., voxel based imaging) with a much higher frequency than in the past (i.e., very fast acquisition). These two latter aspects enable us to develop a new generation of approaches through which address the previously mentioned limitations. The present paper is focused on describing a novel 4D reconstruction method of organs and tissues acquired by high frequency scans. Actually, the proposed method has been primarily designed to reconstruct dynamic 3D models (2D spatial and 1D temporal). Subsequently, it has been extended to reconstruct dynamic 4D models (3D spatial and 1D temporal). This two-steps approach has allowed us to analyze the effectiveness and efficiency of each stage involved in the algorithm. The paper reports an extended experimental session performed on a wide set of 3D CT images which validate the proposed method. Moreover, preliminary tests have been also performed on ad-hoc 4D synthetic images which confirm the success of the results obtained in 3D.