VoronoiTessellations

Python3 script to create Voronoi Tessellations (Mosaic pattern) on images. The script basically does 2D k-means-like clustering of the pixels based on fixed pre-defined cluster centres (which can be set to be random), and averages the RGB values for each cluster group and assigns all the pixels in the group the average value. Further options allow selectively applying average to only one of the RGB channels or exchanging values of the channels randomly (see docs).

This is my first ever open-source repository. :)

Requirements

1. Python3
2. numpy
3. Pillow (PIL fork)
4. tkinter (for GUI. Usually the part of standard Python download)
5. multiprocessing (for the fast method. Part of the standard Python download)

How to Run

The script can be called from the terminal. The simplest run with default options is thus:

python VoronoiMain.py input.jpg output.jpg no_of_clusters

You can run the following command to see help:

python VoronoiMain.py -h
VoronoiMain.py [-h] [--rescale RESCALE] [--border BORDER]
                  [--method METHOD] [--threshold THRESHOLD]
                  [--clusmap CLUSMAP] [--probmap PROBMAP]
                  [--channel {r,g,b,rand,rb,rg,bg,randdual}]
                  [--verbose VERBOSE] [--seed SEED]
                  [--gaussianvars [GAUSSIANVARS [GAUSSIANVARS ...]]]
                  input output cn

positional arguments:
  input                 Input Image file
  output                Output Image file
  cn                    Number of clusters
optional arguments:
  -h, --help            show this help message and exit
  --rescale RESCALE     Rescaling factor for large images
  --border BORDER       Make border [1/0]?
  --method METHOD       fast vs low_mem methods. Default is low_mem.
  --threshold THRESHOLD
                        Only for borders. Threshold distance.
  --clusmap CLUSMAP     Load a specific cluster map as tab-separated text file
  --probmap PROBMAP     Load a 2D probability map for cluster generation
  --channel {r,g,b,rand,rb,rg,bg,randdual}
                        Whether to tessellate along only R,G,B or
                        combinations?
  --verbose VERBOSE     Print progress?[1/0]
  --seed SEED           Seed for PRNG
  --gaussianvars [GAUSSIANVARS [GAUSSIANVARS ...]]
                        Only for gaussian probmap
                        (mx,my,sigmax,sigmay,corr(opt),spacing(opt))

For further details, see the docs.

GUI

This allows an easy way of generating custom cluster maps. Run gui_clusmap.py with the input file and the output file as the arguments. For further details, refer docs.

Performance

For each of the following test, an image of size 1728 x 2304 pixels, and ran on a i7-8th Gen Acer Predator Helios 300 with 16GB RAM. t_clus represents the time in seconds for the cluster computation step, t_avg the time for the averaging step in seconds and the max memory represents the maximum RAM memory used in MB (not very reliable, but approximately true). The seed is 123 wherever applicable. lm stands for low_mem and f stands for fast methods.

Options	t_clus	t_avg	max memory
cn 1000 rescale 2 lm	9	8	30
cn 1000 rescale 2 f	4	4	394
cn 500 border lm	143	20	47
clusmap A lm	48	63	50
clusmap A f	16	8	751
cn 1000 channel r lm	37	13	51
cn 1000 channel r f	13	5	754
cn 500 gaussian lm	28	17	98
cn 500 gaussian f	11	4	930
cn 2000 probmap B lm	50	51	50
cn 2000 probmap B f	14	5	760

Examples

Source:

cn 1500, fast, seed 123

Using a custom clusmap

--channel r and --channel randdual