Research Article

Segmentation of Endothelial Cell Boundaries of Rabbit Aortic Images Using a Machine Learning Approach

Table 2

Comparison of minimum redundancy and maximum relevance in the feature vectors between different training data sets is given in Figure 6. We used four different SVM kernels, but the relative relevance attributed to different features is comparable with all kernels.

Minimum redundancy, maximum relevance
feature order in feature vector from highest to lowest

Training dataset [27 33 3 16 19 32 29 31 21 25 12 31 7 18 26 10 20 1 30 24 9 22 17 11 6 23 15 2 13 8 4 14 5 28]