wasm-color-quant (NEUQUANT)

A WebAssembly-based color quantization module for JavaScript and TypeScript environments. It uses the NEUQUANT algorithm via Rust's color_quant library to reduce RGBA pixel data to a limited color palette, returning both indexed pixels and the color map.

✨ Features

• Fast and compact NEUQUANT implementation in WebAssembly
• Works directly with Uint8Array buffers
• Customizable palette size
• Flat output buffer containing indexed pixels and RGBA palette

📦 Installation

npm install wasm-color-quant

🦀 Rust API (WASM)

quantize_with_palette(pixels: &[u8], sample_factor: i32, palette_size: usize) -> Vec<u8>

Quantizes an RGBA image using NEUQUANT and returns a flat buffer containing the indexed pixel data followed by the RGBA palette.

Parameters

| Name | Type | Description | |----------------|----------|------------------------------------------| | pixels | Uint8Array | Flat RGBA pixel buffer (4 bytes per pixel) | | sample_factor| number | Quality/sample trade-off (e.g. 10) | | palette_size | number | Number of colors to generate (e.g. 256) |

Returns

A Uint8Array where:

• The first pixels.length / 4 bytes are indices.
• The remaining palette_size * 4 bytes are the color palette (RGBA).

🧪 Example (JavaScript)

import init, { quantize_with_palette } from "wasm-color-quant";

await init();

const pixels = new Uint8Array([...]); // Flat RGBA input
const sampleFactor = 10;
const paletteSize = 128;

const result = quantize_with_palette(pixels, sampleFactor, paletteSize);

const indexedPixels = result.slice(0, pixels.length / 4);
const colorPalette = result.slice(pixels.length / 4);

console.log("Indexed:", indexedPixels);
console.log("Palette:", colorPalette);

🧠 Notes

• NEUQUANT performs best on low-frequency, non-photo images like icons and illustrations.
• For photographic images, you may consider alternative quantization algorithms like K-means or Median Cut.

📄 License

MIT

👩‍💻 Author

This module was generated via Rust + wasm-bindgen for modern web-based image processing pipelines.