Table 6 A summary of five publications in terms of their use of segmentation parameter optimization
From: Survey statistics of automated segmentations applied to optical imaging of mammalian cells
Optimized entity | Optimization approach | Segmentation workflow | Reference |
|---|---|---|---|
Intensity threshold, intensity distribution | Otsu technique [43] to minimize intra-class variance | Thresholding→Morphological seeded watershed | [42] |
DIC-based nonnegative-constrained convex objective function minimization→ Thresholding | [44] | ||
Intensity threshold, intensity distribution, geometric characteristics of segmented objects | Find threshold that yields expected size and geometric characteristics | Gaussian filtering→Exponential fit to intensity histogram→Thresholding→ Morphological refinements | [49] |
Thresholding→Morphological refinements | [47] | ||
Intensity distribution, geometric characteristics of segmented objects | Hessian-based filtering and medial axis transform for enhanced intensity-based centroid detection | Iterative non-uniformity correction→Hessian-based filtering→Weighted medial axis transform→Intensity-based centroid detection | [48] |