Computer Vision News - May 2016

Tool Note that the lung segmentation is more accurate with this approach, only the air containing region are selected and the blood vessels were excluded. Once again you can easily browse through the results with the built-in GUI supplied by SimpleITK. Re-run the algorithm with different parameters and observe the new results again. Registration Components in SimpleITK The SimpleITK registration framework supports several optimizer types via the SetMetricAsX() methods. These include: Exhaustive, Nelder-Mead downhill simplex, a.k.a. Amoeba. Variations on gradient descent: (GradientDescent, GradientDescentLineSearch, RegularStepGradientDescent, ConjugateGradientLineSearch) and more. The SimpleITK registration framework supports several metric types via the SetMetricAsX() methods. These include: MeanSquares, Demons, Correlation, ANTSNeighborhoodCorrelation, JointHistogramMutualInformation, MattesMutualInformation, Interpolators. The SimpleITK registration framework supports several interpolators via the SetInterpolator() method, which receives one of the following: sitkNearestNeighbor, sitkLinear, sitkBSpline, sitkGaussian, sitkHammingWindowedSinc, sitkCosineWindowedSinc, sitkWelchWindowedSinc, sitkLanczosWindowedSinc, sitkBlackmanWindowedSinc. It must be said that this process too is interactive and it can be entirely performed in your browser: we only offer a snap view of the interface. For a comprehensive tutorial with additional examples and explanation about the registration process in SimpleITK, we recommend to consult the SimpleITK tutorial workshop held at the last MICCAI conference . COMPUTER VISION NEWS 15