K-Compressor

Reduce the number of colors used in image using k-mean clustering.

Installation

Require go 1.22+

go install github.com/mawngo/kcomp@latest

Usage

compress image

> kcomp .\my-image.jpeg

or compress directory of images

> kcomp .\my-dir

Options

> kcomp -h  
Reduce number of colors used in image

Usage:
  kcomp [files...] [flags]
  
Flags:
  -n, --colors int        Number of colors to use (default 15)
  -o, --out string        Output directory name (default ".")
  -s, --series int        Number of image to generate, series of output with increasing number of colors up util reached --colors parameter [min:1] (default 1)
  -q, --quick             Increase speed in exchange of accuracy
  -w, --overwrite         Overwrite output if exists
  -i, --round int         Maximum number of round before stop adjusting (number of kmeans iterations) (default 100)
  -d, --delta float       Delta threshold of convergence (delta between kmeans old and new centroid’s values) (default 0.005)
  -t, --concurrency int   Maximum number image process at a time [min:1] (default 8)
      --dalgo string      Distance algo for kmeans [EuclideanDistance,EuclideanDistanceSquared] (default "EuclideanDistance")
      --jpeg int          Specify quality of output jpeg compression [0-100] (set to 0 to output png)
      --debug             Enable debug mode
  -h, --help              help for kcomp

Examples

> kcomp .\chika.jpeg --colors=5

10:46PM INF Processing cp=5 round=100 img=chika.jpeg dimension=200x200 format=jpeg
10:46PM INF Compress completed out=chika.100cp5.png took=60.239ms
10:46PM INF Processing completed.

Original	5 Colors	4 Colors