| Method | Description | Related references
|
| Thresholding | Determining threshold and then the pixels are divided into groups based on that criterion. It include bilevel and multithresholding | Histogram thresholding ([90, 91, 109, 110, 223, 224]) Adaptive thresholding ([61, 88, 111, 112, 124, 126, 200, 225]) |
| Color-based segmentation algorithms | Segmentation based on color discrimination. Include principle component transform/spherical coordinate transform | [134, 226–230] |
| Discontinuity-based segmentation | Detection of lesion edges using active contours/radial search techniques/zero crossing of Laplacian of Gaussian (LoG) | Active contours ([58, 62, 64, 88, 231–233]) Radial search ([115, 234, 235]) LoG ([117, 163, 236, 237]) |
| Region-based segmentation | Splitting the image into smaller components then merging subimages which are adjacent and similar in some sense. It includes Statistical region merging, multiscale region growing, and morphological flooding | Split and merge ([238, 239]) SRM ([59, 70, 92, 112]) Multi-scale ([118, 196]) Morphological flooding ([79]) |
| Soft computing | Methods involve the classification of pixels using soft computing techniques including neural networks, fuzzy logic, and evolutionary computation | Fuzzy logic ([60, 76, 80, 85, 140, 157, 200, 240]) Neural Network ([177, 241]) Optimization algorithms ([241–243]) |
|
|