To recognize the characteristics of coiony images, an important step is to segment colony images (delineate colonies).
Therefore, an algorithm based on kernel spatial FCM (fuzzy c-means) is studied for colony images, presented in this paper.
When conventional fuzzy c-means clustering algorithm is used to segment colony images, spatial information is not
considered, and Euclidean distance calculation in such an algorithm is not robust. In this paper, we consider the spatial
information when colony images are deal with, by using MCF. By using Mercer kernel functions, image pixels are mapped
from the original space into a higher dimensional feature space. We can perform c-means clustering efficiently in the feature
space for the kernel functions, which can induce robust distance measures while the computational complexity is low. We
conduct some experiments on colony images by using the new algorithm. The results show that the studied algorithm is
suitable and robust for colony images segmentation.