Noise based in speckle

This synthetic dataset or phantom consists of 3 jpg format databases, in the two-dimensional (2-D) domain, which are identified as follows: DB1: Ground Truth DB2: Speckle noise with zero mean and 0.005 standard deviation DB3: Speckle noise with zero mean and 0.05 standard deviation Considering a real cardiac CT database, images with information about both the left ventricular cavity (LVC) and the left ventricle wall (LVW) were selected. Then, the gray levels from these images were used to construct each of the aforementioned databases. This real database was acquired with a GE Bright Speed Elite 16-slice scanner and was donated by the Laboratory of Traitement signal et de l’image, belonging to Rennes I University, France.Thus, to generate all synthetic DBs, specific gray levels were considered, and the myocardium and LV wall was simulated by an inner cone and an outer cone, respectively. The spatial resolution of both cones was 256 × 256 × 50 voxels. The value 256 is equivalent to 50% of the spatial dimension of a typical cardiac tomography layer, whereas the value 50 corresponds to approximately 20% of the number of layers of the actual data analyzed.All the databases were built in a set of concentric circular circles or layers, which were conditioned to contain densitometric information obtained by considering the average gray levels (GL) exhibited by a real cardiac computed tomography database (CT_DB). To introduce in the databases the information related to the GLs, each of the generated circular layers was processed with a filling technique based on a region growing algorithm