Figure 8

Machine segmentation performances on Drosophila brain segmentation. Performance evaluation of six MS algorithms segmenting Drosophila brain stacks (n = 5). All algorithms calculate a single global threshold. When that threshold is smaller than that of GT, there will not be any false positive voxel which leads to 100% precision (p = 1, orange) in case of Yen, Rényi, IsoData and Otsu methods. When the MS threshold is calculated to be larger than that of GT, there will not be any false negative voxel resulting in 100% recall (r = 1, yellow) in the case of the Li and Huang methods. F-measure values are represented by the blue bars. Error bars represent standard deviation.