Bioleptic

Wavelet-based signal compression for physiological data (ECG, PPG, accelerometry).

Algorithm

1. Non-finite substitution (NaN → 0, ±inf → 0/1)
2. Mean-centering + range normalization
3. Multi-level DWT (CDF 5/3 or CDF 9/7, 5 levels)
4. Quantization to i16 with configurable scale
5. Detail coefficient thresholding
6. Entropy coding with deflate

Rust

let compressed = compress(&signal, CompressionOptions::default())?;
let recovered  = decompress(&compressed)?;

Python

pip install bioleptic-py

from bioleptic import compress_signal, decompress_signal, BiolpCompressionOptions

compressed = compress_signal(signal, BiolpCompressionOptions("cdf97", 11, "low"))
recovered  = decompress_signal(compressed)

JavaScript

npm install bioleptic-js

import {
    BiolpCompressionMethod,
    BiolpCompressionOptions, BiolpCutoffLevel,
    BiolpQuantizationScale,
    compress_signal,
    decompress_signal
} from "bioleptic-js";

const signal = new Float32Array([1.0, 2.0, 3.0, 4.0, 5.0, 1.0, 2.0, 3.0, 4.0, 5.0, 1.0, 2.0, 3.0, 4.0, 5.0]);

const options = new BiolpCompressionOptions(
    BiolpCompressionMethod.Cdf97,
    BiolpQuantizationScale.S11,
    BiolpCutoffLevel.Low,
);

const compressed = compress_signal(signal, options);  // Uint8Array
const recovered = decompress_signal(compressed);     // Float32Array
console.info("Recovered signal", recovered);

This project is licensed under either of

• BSD-3-Clause License (see LICENSE)
• Apache License, Version 2.0 (see LICENSE)

at your option.