fast-skimage 0.3.1

Fast and easy image processing using an Image class based on the scikit-image, numpy and matplotlib libraries.

fast-skimage

Welcome to fast-skimage, an image acquisition and processing library. This powerful library offers a wide range of tools for advanced image manipulation and analysis, wrapped up in the accessible Image class.

Features

• Advanced Manipulation: Apply complex operations like adding watermarks, noise detection, auto-enhancement, and saturation increase with simple method calls.
• Filtering and Thresholding: Includes mean, median filtering, Otsu's thresholding, and custom thresholding methods for image segmentation and noise reduction.
• Fourier Transforms: Utilize Fourier-based methods for reducing image dithering and other artifacts.
• Histogram Operations: Equalize and stretch image histograms to improve contrast and visibility.
• Texture Analysis: Perform texture segmentation using a variety of descriptors.
• Small Image Library: 7 various pictures for testing are provided with the package (see section "Image Library" below).

Getting Started

1. Installation: Clone the repository or download the Image class module to your project.
2. Dependencies: Ensure all dependencies such as numpy, matplotlib, scikit-image, and PyWavelets are installed.
3. Usage: Import the Image class from the module and instantiate it with the path to your image or a NumPy array.

Example

from fast_skimage import Image
from fast_skimage import etretat
from skimage.data import immunohistochemistry

img = Image("Pictures/camera.jpg")  # Load an image with path...
img2 = Image(immunohistochemistry())  # ... or numpy array ...
colored_image_array = etretat() # ... or a library image.
img3 = Image(colored_image_array.get())

img2.auto_enhance()  # Apply auto-enhancement
img3.auto_enhance()

img3.show(subplots=(1, 2, 1), size=12)  # Display the result
img2.show(subplots=(1, 2, 2), title='Immunochemistry Image')

img.show(size=(12, 6), type_of_plot='hist', axis=True)  # Plot histogram

Image Library

A small image library is provided along with the Image class. These can be manually extracted with the following lines:

from fast_skimage import image_name
image_array = image_name()
image = Image(image_array.get())

Note that all images listed below come from the INFO-H500 course of Prof. Olivier Debeir at ULB (Université Libre de Bruxelles).

Grayscale Noisy Image

• fast-skimage.astronaut_noisy

Grayscale Clean Images

• fast-skimage.camera
• fast-skimage.walking

Grayscale Clean Watermark

• fast-skimage.watermark (the ULB logo)

Colored Clean Images

• fast-skimage.etretat
• fast-skimage.nyc
• fast-skimage.zebra

Collection Library

The Collection class allows you to handle multiple images (Image objects) at the same time. For instance:

from fast_skimage.Collection import Collection

# Loads all images as Image objects in a single Collection object
tr_glioma = Collection(directory_path="Training/glioma", name="Glioma (training)", verbose=1)
keys, images = tr_glioma.images.items()
# Extract all image pixels back into a Pandas dataframe
glioma = tr_glioma.extract_images(label="Glioma", to_df=True)

Documentation

Refer to the in-line comments and method docstrings for detailed usage of each feature.

Contribution

Contributions are welcome! Feel free to submit pull requests, suggest features, or report bugs.

License

This library is distributed under the MIT license. See LICENSE for more information.

Contact

• Author: Alexandre Le Mercier
• Date: November 21, 2023 (last update on February 11, 2024)
• Email: [e-mail]
• LinkedIn: Alexandre Le Mercier

Happy Image Processing!