torch-image-lerp

Linear 2D/3D image interpolation and gridding in PyTorch.

Why?

This package provides a simple, consistent API for

• sampling from 2D/3D images (sample_image_2d()/sample_image_3d())
• inserting values into 2D/3D images (insert_into_image_2d(), insert_into_image_3d)

Operations are differentiable and sampling from complex valued images is supported.

Installation

pip install torch-image-lerp

Usage

Sample from image

import torch
import numpy as np
from torch_image_lerp import sample_image_2d

image = torch.rand((28, 28))

# make an arbitrary stack (..., 2) of 2d coords
coords = torch.tensor(np.random.uniform(low=0, high=27, size=(6, 7, 8, 2))).float()

# sampling returns a (6, 7, 8) array of samples obtained by linear interpolation
samples = sample_image_2d(image=image, coordinates=coords)

The API is identical for 3D but takes (..., 3) coordinates and a (d, h, w) image.

Insert into image

import torch
import numpy as np
from torch_image_lerp import insert_into_image_2d

image = torch.zeros((28, 28))

# make an arbitrary stack (..., 2) of 2d coords
coords = torch.tensor(np.random.uniform(low=0, high=27, size=(3, 4, 2)))

# generate random values to place at coords
values = torch.rand(size=(3, 4))

# sampling returns a (6, 7, 8) array of samples obtained by linear interpolation
samples = insert_into_image_2d(values, image=image, coordinates=coords)

The API is identical for 3D but takes (..., 3) coordinates and a (d, h, w) image.