navalmartin-mir-vision-utils 0.0.4

Various utilities for working with images in Python 3. for the mir project

navalmartin_mir_vision_utils

Various utilities for working with images in the mir project.

Acknowledgements

The project incorporates the following repositories

• image-quality https://github.com/ocampor/image-quality (for BRISQUE)
• imutils: https://github.com/PyImageSearch/imutils (for various utilities with OpenCV)

Dependencies

• torch
• torchvision
• numpy
• dataclasses
• Pillow
• matplotlib
• opencv-python
• scipy
• scikit-image
• libsvm

Installation

Installing the utilities via pip

pip install navalmartin-mir-vision-utils

For a specific version use

pip install navalmartin-mir-vision-utils==x.x.x

You can uninstall the project via

pip3 uninstall navalmartin-mir-vision-utils

How to use

Below are some use-case samples. You can find more in the examples.

Using image_utils

from pathlib import Path
from navalmartin_mir_vision_utils.image_utils import (is_valid_pil_image_file, get_pil_image_size,
                                                      get_img_files)


if __name__ == '__main__':

    image = is_valid_pil_image_file(image=Path("/home/alex/qi3/mir-engine/datasets/cracks_v_3_id_8/train/cracked/img_9_9.jpg"))
    if image is not None:
        print("The provided image is OK")
        image_size = get_pil_image_size(image=image)
        print(f"Image size is {image_size}")
    else:
        print("The provided image is NOT OK")

    base_path = Path("/home/alex/qi3/mir-engine/datasets/cracks_v_3_id_8/train/cracked/")
    image_files = get_img_files(base_path=base_path)
    print(f"There are {len(image_files)} in {base_path}")

Using image_transformers

from pathlib import Path
from navalmartin_mir_vision_utils.image_transformers import pil_image_to_bytes_string
from navalmartin_mir_vision_utils.image_utils import load_img
from navalmartin_mir_vision_utils.image_enums import ImageLoadersEnumType
from navalmartin_mir_vision_utils.image_utils import is_valid_pil_image_from_bytes_string
from navalmartin_mir_vision_utils.image_utils import show_pil_image

if __name__ == '__main__':

    image_path = Path("/home/alex/qi3/mir-engine/datasets/cracks_v_3_id_8/train/cracked/img_9_9.jpg")
    image = load_img(path=image_path, loader=ImageLoadersEnumType.PIL)

    show_pil_image(image=image)

    image_bytes = pil_image_to_bytes_string(image=image)
    image = is_valid_pil_image_from_bytes_string(image_byte_string=image_bytes)
    show_pil_image(image=image)