Figure 2

Comparison of k-means clustering and IMFSegNet. The Plot show the averaged fat amount of the nPCA and cdPCA subgroups of four control sheep (n = 4) of the PCA muscle without RLN-damage (nPCA) and with RLN-damage (cdPCA), which was determined by k-means (yellow) and IMFSegNet (green). The measured fat amount differs significantly between k-means and IMFSegNet in the nPCA (p-value = 0.04, d = 1.82) and in the cdPCA (p-value = 0.01, d = 3.54). While the IMFSegNet has no significant differences with a medium effect (p-value = 0.40, d = 0.65), k-means clustering differs significantly with a large effect (p-value = 0.02, d = 2.39). White dots denote mean values (with dot) and black lines denote the median (middle line).