Basic Cardinalities For two net partitions (segmentations) Sg and St, the confusion matrix consists of the four common cardinalities that reflect the overlap between the two partitions, namely TP, FP, FN and TN. These cardinalities see for each pair of subsets i∈Sg and I∈t the sum of the correspondence between them. It was the deputy who explained that the restrictions were abandoned because The Exodus Tohas plan of Central broke to exist peacefully inside the city. He predicts that the plan will fail dangerously, depriving Killy and Cibo of any chance of finding a networked terminal gene in Toha. The research was funded under the European Union`s Seventh Framework Programme (PC7/2007-2013) under grant agreement 318068 (VISCERAL). The Cohen Kappa coefficient (KAP) proposed in  is a measure of the match between two samples. As an advantage over other measures, KAP takes into account the fortuitous agreement, making it more robust. KAP is proposed by Gerig et al.  a tool (Valmet) to evaluate the segmentation of medical volume. In this tool, only five metrics are implemented. There are important metrics, such as theoretical information counters, as well as certain statistical metrics such as mahalanobis ablation and chance-correcting metrics such as kappa and adapted rand index, which are not implemented in the Valmet assessment tool. In addition, this tool does not provide support for fuzzy segmentation. Itk Library2 offers a software layer that supports medical imaging tasks, including segmentation and recording.
The ITK library provides evaluation metrics that are usually based on distance transformation filters . However, this implementation has the following flaws: First, the ITK library does not implement all the relevant measures necessary to evaluate medical segmentation. Second, because most metrics are based on distance-processing filters, they are sensitive to increasing the size of the volume grids in terms of speed and memory used. One way to reduce this effect is to use the delimit cube (scene) [6, 10], i.e. to exclude from the calculation the smallest cube with the two segments, i.e. all the background voxels that are not in the template. However, there are some defects with the ceiling, first the delimitation cube remains large when the segments are large, far from each other or there are fugues away from the segments; Second, the delimit court affects the results of metrics that depend on actual negatives.