Automatic Segmentation of Pathological Glomerular Basement Membrane in Transmission Electron Microscopy Images with Random Forest Stacks
Algorithm 1
(1)
Coarse segmentation results with an obvious error GBM area are eliminated on the basis of the preset threshold .
(2)
The remaining coarse segmentation results are used to obtain the probability map and the binary mask as shown in Sections 2.4.1 and 2.4.2.
(3)
The Jaccard similarity of and each are calculated. If the similarity is less than a preset threshold , the corresponding is abandoned. The remaining segmentation results are used to obtain a new probability map and the binary mask .
(4)
The Jaccard similarity of and is calculated. If , let and , and step (3) is repeated until before they exit the loop.
(5)
After the loop ends, the candidate result equal to is obtained.
