By applying the hybrid refinement scheme on a well-known corpus, significant improvements of segmentation results in terms of segmentation accuracy with different tolerances, mean absolute error (MAE), and root-mean-square error (RMSE) can be observed.
Although there have been many significant object-oriented segmentation algorithms in remote sensing image segmentation using multispectral information [3 5], only a few ones can be effectively applied in the segmentation of high-resolution remote sensing images.
The absence of a clear boundary between these structures poses a significant challenge for segmentation.
The approach is compared to other two significant natural scene segmentation methods, achieving better results in a global evaluation.
Experimental results demonstrate the effectiveness of the proposed system on real world images, and indicate significant improvement in segmentation quality over the base system.
Therefore, a significant number of segmentation errors can be expected when the automated analysis is performed in patients with macular holes.
However, we could not detect any significant mismatches in segmentation of the subcortical structures performed by the algorithm (Figs 1 and 2 are a 2D-represenations of segmented subcortical structures by FIRST).
A significant effect of segmentation was found on MD (F = 29.96, df = 1, 29, p < .001) but not FA (F = 1.80, df = 1, 29, p = .191), with MD being higher in the ventral compared to the dorsal segment.
In addition to microstructural markers, a significant effect of segmentation on streamline count was found (F = 25.04, df = 1, 29, p < .001), with a larger streamline count in dorsal segments compared to ventral segments (Fig. 4), independent of hemisphere (F = 4.09, df = 1, 29, p = .052).052
