Source code for medical_image.process.frequency

import torch

from medical_image.data.image import Image, requires_loaded
from medical_image.utils.device import resolve_device




class FrequencyOperations:


    @staticmethod
    @requires_loaded
    def fft(image: Image, output: Image, device=None) -> Image:
        """
        Computes the 2-dimensional Fast Fourier Transform (FFT) of an image.

        Args:
            image: Input image.
            output: Output image to store the complex FFT result.
            device: Device to perform computation on (None = infer from image).

        Returns:
            The output Image.
        """
        device = resolve_device(image, explicit=device)
        img = image.pixel_data.to(device).float()
        fft_result = torch.fft.fft2(img)
        output.pixel_data = fft_result.to(device)
        return output




    @staticmethod
    @requires_loaded
    def inverse_fft(image: Image, output: Image, device=None) -> Image:
        """
        Computes the inverse 2-dimensional Fast Fourier Transform (IFFT) of an image.

        Args:
            image: Input image in the frequency domain (complex tensor).
            output: Output image to store the inverse FFT result.
            device: Device to perform computation on (None = infer from image).

        Returns:
            The output Image.
        """
        device = resolve_device(image, explicit=device)
        img = image.pixel_data.to(device)
        ifft_result = torch.fft.ifft2(img)
        output.pixel_data = ifft_result.to(device)
        return output