tomodrgn.fft.iht2_center_torch

iht2_center_torch(possibly_batched_imgs: Tensor) → Tensor

Batch-vectorized 2-D inverse DHT of an even-box sized image(s) as torch tensor By definition, is equivalent to forward DHT up to normalization scale factor (of N samples, typically D*D here) Note: torch maintains default input dtype torch.float32 to output dtype torch.float32

Parameters:

possibly_batched_imgs – torch.tensor of images, shape ([N], D, D)

Returns:

torch.tensor of images DFT, shape ([N], D, D)