Module resampling 

SIMD-accelerated resampling operations

This module provides high-performance image resampling using SIMD instructions for bilinear and bicubic interpolation. The implementations use cache-friendly blocking strategies for optimal performance on large rasters.

Supported Methods

• Bilinear: Fast, smooth interpolation (2x2 kernel)
• Bicubic: High-quality interpolation (4x4 kernel)
• Batch Processing: Process multiple pixels in parallel with SIMD

Performance

Expected speedup over scalar: 2-4x depending on interpolation method

Example

use oxigdal_algorithms::simd::resampling::bilinear_f32;

let src = vec![1.0_f32; 100 * 100];
let mut dst = vec![0.0_f32; 50 * 50];

bilinear_f32(&src, 100, 100, &mut dst, 50, 50)?;

Functions

bicubic_f32

Bicubic interpolation with SIMD optimization

bilinear_f32

Bilinear interpolation with SIMD optimization

downsample_average_f32

Downsample using area averaging (for antialiasing)

nearest_f32

Nearest neighbor resampling (fast, no interpolation)