Module image_ops 

Image processing operations using 2D FFTs.

This module provides common image processing tasks that leverage 2D FFTs for efficient computation:

• Convolution and correlation
• Spatial filtering (low-pass, high-pass, band-pass)
• Edge detection in frequency domain
• Image sharpening and blurring

Performance

FFT-based convolution is faster than spatial convolution for large kernels (typically > 7x7), due to the convolution theorem: multiplication in the frequency domain is equivalent to convolution in the spatial domain.

Examples

use spectrograms::image_ops::{gaussian_kernel_2d, convolve_fft};
use spectrograms::nzu;
use ndarray::Array2;

// Create a 256x256 image
let image = Array2::<f64>::zeros((256, 256));

// Apply Gaussian blur
let kernel = gaussian_kernel_2d(nzu!(9), 2.0).unwrap();
let blurred = convolve_fft(&image.view(), &kernel.view()).unwrap();

Functions

bandpass_filter

Apply band-pass filter (keep frequencies in a specific range).

convolve_fft

Convolve 2D image with kernel using FFT (faster for large kernels).

detect_edges_fft

Detect edges in an image using high-pass filtering.

gaussian_kernel_2d

Create 2D Gaussian kernel for blurring.

highpass_filter

Apply high-pass filter (suppress low frequencies).

lowpass_filter

Apply low-pass filter (suppress high frequencies).

sharpen_fft

Sharpen an image by enhancing high frequencies.