Module threshold 

Adaptive thresholding implementation. Adaptive thresholding and image binarization.

This module provides the first stage of the detection pipeline, converting grayscale input into binary images while adapting to local lighting conditions.

It features two primary implementations:

1. Tile-based Thresholding: Blazing fast approach using local tile stats.
2. Integral Image Thresholding: Per-pixel adaptive thresholding for small features.

Structs

ThresholdEngine

Adaptive thresholding engine using tile-based stats.

TileStats

Statistics for a single tile.

Functions

adaptive_threshold_gradient_window

Apply per-pixel adaptive threshold with gradient-based window sizing.

adaptive_threshold_integral

Apply per-pixel adaptive threshold using integral image.

apply_adaptive_threshold_fast

Fast adaptive threshold combining integral image approach with SIMD.

apply_adaptive_threshold_with_params

Adaptive threshold with custom parameters.

compute_integral_image

Compute integral image (cumulative sum) for fast box filter computation.

compute_threshold_map

Compute a map of local mean values.