tiler 0.2.2

Fast N-dimensional numpy array tiling and merging with overlapping, padding and tapering support

tiler

Github repository | Github issues | Documentation

---

⚠️ Please note: work in progress, things will change and/or break! ⚠️

---

This python package provides functions for tiling/patching and subsequent merging of NumPy arrays.

Such tiling is often required for various heavy image-processing tasks such as semantic segmentation in deep learning, especially in domains where images do not fit into GPU memory (e.g., hyperspectral satellite images, whole slide images, videos, tomography data).

Features

• N-dimensional (note: currently tile shape must have the same number of dimensions as the array)
• Optional in-place tiling
• Optional channel dimension, dimension that is not tiled
• Optional tile batching
• Tile overlapping
• Access individual tiles with iterator or a getter
• Tile merging, with optional window functions/tapering

Quick start

This is an example of basic functionality.
You can find more examples in examples.
For more Tiler and Merger functionality, please check documentation.

import numpy as np
from tiler import Tiler, Merger

image = np.random.random((3, 1920, 1080))

# Setup tiling parameters
tiler = Tiler(data_shape=image.shape,
              tile_shape=(3, 250, 250),
              channel_dimension=0)

## Access tiles:
# 1. with an iterator
for tile_id, tile in tiler.iterate(image):
   print(f'Tile {tile_id} out of {len(tiler)} tiles.')
# 1b. the iterator can also be accessed through __call__
for tile_id, tile in tiler(image):
   print(f'Tile {tile_id} out of {len(tiler)} tiles.')
# 2. individually
tile_3 = tiler.get_tile(image, 3)
# 3. in batches
tiles_in_batches = [batch for _, batch in tiler(image, batch_size=10)]

# Setup merging parameters
merger = Merger(tiler)

## Merge tiles:
# 1. one by one
for tile_id, tile in tiler(image):
   merger.add(tile_id, some_processing_fn(tile))
# 2. in batches
merger.reset()
for batch_id, batch in tiler(image, batch_size=10):
   merger.add_batch(batch_id, 10, batch)

# Final merging: applies tapering and optional unpadding
final_image = merger.merge(unpad=True)  # (3, 1920, 1080)

Installation

The latest release is available through pip:

pip install tiler

Alternatively, you can clone the repository and install it manually:

git clone [e-mail]:the-lay/tiler.git
cd tiler
pip install

Roadmap

• Easy generation of tiling for a specific window in mind (i.e. so that every element has the window weight sum of 1.0)
• Add border windows generation (like in Pielawski et. al - see references))
• PyTorch Tensors support
  • merging on GPU like in pytorch-toolbelt?
• More examples
• Implement windows functions and remove scipy dependency (we need only a couple of functions that generate windows)
• PyTorch Dataset class convenience wrapper?
• Arbitrary sized tiles (m-dim window over n-dim array, m <= n)?
  • Some leads here
• Optional augmentation modes for smoother segmentations?
  • D4 rotation group
  • Mirroring
• Benchmark with plain for loops, determine overhead

Motivation & other packages

I work on semantic segmentation of patched 3D data and I often found myself reusing tiling functions that I wrote for the previous projects. No existing libraries listed below fit my use case, so that's why I wrote this library.

However, other libraries might fit you better:

• vfdev-5/ImageTilingUtils

  • Minimalistic image reader agnostic 2D tiling tools

• BloodAxe/pytorch-toolbelt

  • Powerful PyTorch toolset that has 2D image tiling and on-GPU merger

• Vooban/Smoothly-Blend-Image-Patches

  • Mirroring and D4 rotations data (8-fold) augmentation with squared spline window function for 2D images

• samdobson/image_slicer

  • Slicing and merging 2D image into N equally sized tiles

• dovahcrow/patchify.py

  • Tile and merge 2D, 3D images defined by tile shapes and step between tiles

• Do you know any other similar packages?

  • Please make a PR or open a new issue.

Moreover, some related approaches have been described in the literature:

• Introducing Hann windows for reducing edge-effects in patch-based image segmentation, Pielawski and Wählby, March 2020