Module threshold 

SIMD-accelerated thresholding operations

This module provides high-performance image thresholding and binarization using SIMD instructions. Thresholding is fundamental for segmentation, feature extraction, and image preprocessing.

Supported Operations

• Binary Thresholding: Simple threshold with two output values
• Adaptive Thresholding: Local threshold based on neighborhood
• Otsu’s Method: Automatic threshold selection
• Multi-level Thresholding: Multiple threshold values
• Range Thresholding: Keep values within a range
• Hysteresis Thresholding: Two-level threshold with connectivity

Performance

Expected speedup over scalar: 6-12x for thresholding operations

Example

use oxigdal_algorithms::simd::threshold::{binary_threshold, otsu_threshold};
use oxigdal_algorithms::error::Result;

fn example() -> Result<()> {
    let data = vec![128u8; 1000];
    let mut output = vec![0u8; 1000];

    binary_threshold(&data, &mut output, 100, 255, 0)?;
    Ok(())
}

Functions

adaptive_threshold_gaussian

Apply adaptive threshold using Gaussian-weighted mean

adaptive_threshold_mean

Apply adaptive threshold using local mean

binary_threshold

Binary threshold with custom output values

hysteresis_threshold

Hysteresis thresholding (two-level threshold with connectivity)

multi_threshold

Multi-level thresholding

otsu_threshold

Calculate optimal threshold using Otsu’s method

threshold_range

Range threshold - keep values within [low, high]

threshold_to_zero

Binary threshold to zero

threshold_truncate

Truncate threshold - cap values at threshold