Updated 09 May 2019
c-means and fuzzy c-means clustering are two very popular image segmentation algorithms. While their implementation is straightforward, if realized naively it will lead to substantial overhead in execution time and memory consumption. Although these deficiencies could be ignored for small 2D images they become more noticeable for large 3D datasets. This submission is intended to provide an efficient implementation of these algorithms for segmenting N-dimensional grayscale images. The computational efficiency is achieved by using the histogram of the image intensities during the clustering process instead of the raw image data. Finally, since the algorithms are implemented from scratch there are no dependencies on any auxiliary toolboxes.
For a quick demonstration of how to use the functions, run the attached 'DemoFCM' file.
If you run into problems using submitted functions, please report them here:
Anton Semechko (2020). Fast fuzzy c-means image segmentation (https://github.com/AntonSemechko/Fast-Fuzzy-C-Means-Segmentation), GitHub. Retrieved .
Which method? Image segmentation via clustering of pixel/voxel intensities?
is there any research paper related to this method
Sir...i want my output image as gray image not color how to modify this code..kindly suggest me..plz..
hello sir...great job....i want to embedd this code with MRF..kindly guide me...as i want to give ffcm as input for mrf for initial segmentation...when i do this..it shows error with reshape function..y so? plz help....waiting for reply sir...
May I know what is the algorithm used? Do you have any article? Thanks
if i'd like to do it based on color values and my input image is COLOR image... then? Plz guide
This is a very good replacement for the kmeans function, and it does not require any toolbox, and works N-D.
- title typo
- updated submission description
migrated to GitHub
Included a function that transforms 1D fuzzy memberships to fuzzy membership maps.