https://github.com/delthas/octreequant

A simple library for converting images to paletted images (color quantization) fast.
https://github.com/delthas/octreequant

Last synced: 2 days ago
JSON representation

A simple library for converting images to paletted images (color quantization) fast.

• Host: GitHub
• URL: https://github.com/delthas/octreequant
• Owner: delthas
• License: mit
• Created: 2024-07-05T22:34:01.000Z (4 months ago)
• Default Branch: master
• Last Pushed: 2024-07-24T15:23:56.000Z (4 months ago)
• Last Synced: 2024-07-24T17:55:31.586Z (4 months ago)
• Language: Go
• Size: 1.07 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# octreequant [![GoDoc](https://godocs.io/github.com/delthas/octreequant?status.svg)](https://godocs.io/github.com/delthas/octreequant)

A simple library for converting images to paletted images (color quantization) fast.

Quantizing an image takes a substantial amount of time, in the order of a second, so using an optimized algorithm/library matters. This library uses the fastest mainstream quantization algorithm.

Main quantization algorithms:

- Slowest: Generating a palette, then "**brute-forcing**" all the possible palettes for each pixel (what Go's draw.Draw does)

- Fast: Generating clusters of pixels, then cutting the biggest in half on their **median** element until the image has <= N clusters (what https://github.com/soniakeys/quant does)

- Fastest: Adding all the pixel colors to an **octree** (tree with 8 children per node), then pruning the tree bottom-up until it has <= N leafs (**this repo**)

In practice, this library can be typically used to render SIXEL images fast.

On a typical system, encoding a 1000x1000 image takes:

- 2s with the brute-force algorithm

- 0.8s with the median algorithm

- 0.3s with this repo's algorithm

## Usage

```

go get github.com/delthas/octreequant@master

```

The API is well-documented in its [![GoDoc](https://godocs.io/github.com/delthas/octreequant?status.svg)](https://godocs.io/github.com/delthas/octreequant)

To convert an image to a paletted image with a 256-color palette:

```

var in image.Image // = ...

out := octreequant.Paletted(in, 256)

```

## Example

Before quantization:

![Original image](landscape.png)

After quantization:

![Quantized image](landscape.quantized.png)

## License

MIT