The final goal of segmentation is generally to assign a label to every pixel of the input image such that pixels with the same label share certain visual characteristics (objects or regions). This segmented image can be then processed by analysis tools to make measurements on its components (see section Image Analysis).
This category contains several groups of engines that allow for performing image segmentation :
- Binarization: to create a binary image where pixels/voxels to analyze are set to intensity 1 (objects) and other, set to 0, are considered as background
- Labelization: to create a label image where each pixels belonging to the same entity are set to a same intensity (its label).
- Discrete geometry : to simplify edges of objects making a polygonal approximation.
- Watershed: to segment an image using the watershed algorithm.