# Deskew
Note: Skew is measured in degrees. Deskewing is a process whereby skew is removed by rotating an image by the same amount as its skew but in the opposite direction. This results in a horizontally and vertically aligned image where the text runs across the page rather than at an angle.
The return angle is between -45 and 45 degrees to don't arbitrary change the image orientation.
By using the library you can set the argument `angle_pm_90` to `True` to have an angle between -90 and 90 degrees.
## Skew detection and correction in images containing text
| Image with skew | Image after deskew |
| ---------------------------------------------------- | ------------------------------------------------------------------ |
|  |  |
## Installation
You can install deskew directly from pypi directly using the following comment
```bash
python3 -m pip install deskew
```
Or to upgrade to newer version
```bash
python3 -m pip install -U deskew
```
## Cli usage
Get the skew angle:
```bash
deskew input.png
```
Deskew an image:
```bash
deskew --output output.png input.png
```
## Lib usage
With scikit-image:
```python
import numpy as np
from skimage import io
from skimage.color import rgb2gray
from skimage.transform import rotate
from deskew import determine_skew
image = io.imread('input.png')
grayscale = rgb2gray(image)
angle = determine_skew(grayscale)
rotated = rotate(image, angle, resize=True) * 255
io.imsave('output.png', rotated.astype(np.uint8))
```
With OpenCV:
```python
import math
from typing import Tuple, Union
import cv2
import numpy as np
from deskew import determine_skew
def rotate(
image: np.ndarray, angle: float, background: Union[int, Tuple[int, int, int]]
) -> np.ndarray:
old_width, old_height = image.shape[:2]
angle_radian = math.radians(angle)
width = abs(np.sin(angle_radian) * old_height) + abs(np.cos(angle_radian) * old_width)
height = abs(np.sin(angle_radian) * old_width) + abs(np.cos(angle_radian) * old_height)
image_center = tuple(np.array(image.shape[1::-1]) / 2)
rot_mat = cv2.getRotationMatrix2D(image_center, angle, 1.0)
rot_mat[1, 2] += (width - old_width) / 2
rot_mat[0, 2] += (height - old_height) / 2
return cv2.warpAffine(image, rot_mat, (int(round(height)), int(round(width))), borderValue=background)
image = cv2.imread('input.png')
grayscale = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
angle = determine_skew(grayscale)
rotated = rotate(image, angle, (0, 0, 0))
cv2.imwrite('output.png', rotated)
```
## Debug images
If you get wrong skew angle you can generate debug images, that can help you to tune the skewing detection.
If you install deskew with `pip install deskew[debug_images]` you can get some debug images used for
the skew detection with the function `determine_skew_debug_images`.
To start the investigation you should first increase the `num_peaks` (default `20`) and use
the `determine_skew_debug_images` function.
Then you can try to tune the following arguments `num_peaks`, `angle_pm_90`, `min_angle`, `max_angle`,
`min_deviation` and eventually `sigma`.
Inspired by Alyn: https://github.com/kakul/Alyn
## Contributing
Install the pre-commit hooks:
```bash
pip install pre-commit
pre-commit install --allow-missing-config
```
Raw data
{
"_id": null,
"home_page": "https://github.com/sbrunner/deskew",
"name": "deskew",
"maintainer": "",
"docs_url": null,
"requires_python": ">=3.9",
"maintainer_email": "",
"keywords": "",
"author": "St\u00e9phane Brunner",
"author_email": "stephane.brunner@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/2f/dd/dc20f529fc1246e52e851a79ae718951134dda366c070e48c08c03654c3c/deskew-1.5.1.tar.gz",
"platform": null,
"description": "# Deskew\n\nNote: Skew is measured in degrees. Deskewing is a process whereby skew is removed by rotating an image by the same amount as its skew but in the opposite direction. This results in a horizontally and vertically aligned image where the text runs across the page rather than at an angle.\n\nThe return angle is between -45 and 45 degrees to don't arbitrary change the image orientation.\n\nBy using the library you can set the argument `angle_pm_90` to `True` to have an angle between -90 and 90 degrees.\n\n## Skew detection and correction in images containing text\n\n| Image with skew | Image after deskew |\n| ---------------------------------------------------- | ------------------------------------------------------------------ |\n|  |  |\n\n## Installation\n\nYou can install deskew directly from pypi directly using the following comment\n\n```bash\npython3 -m pip install deskew\n```\n\nOr to upgrade to newer version\n\n```bash\npython3 -m pip install -U deskew\n```\n\n## Cli usage\n\nGet the skew angle:\n\n```bash\ndeskew input.png\n```\n\nDeskew an image:\n\n```bash\ndeskew --output output.png input.png\n```\n\n## Lib usage\n\nWith scikit-image:\n\n```python\nimport numpy as np\nfrom skimage import io\nfrom skimage.color import rgb2gray\nfrom skimage.transform import rotate\n\nfrom deskew import determine_skew\n\nimage = io.imread('input.png')\ngrayscale = rgb2gray(image)\nangle = determine_skew(grayscale)\nrotated = rotate(image, angle, resize=True) * 255\nio.imsave('output.png', rotated.astype(np.uint8))\n```\n\nWith OpenCV:\n\n```python\nimport math\nfrom typing import Tuple, Union\n\nimport cv2\nimport numpy as np\n\nfrom deskew import determine_skew\n\n\ndef rotate(\n image: np.ndarray, angle: float, background: Union[int, Tuple[int, int, int]]\n) -> np.ndarray:\n old_width, old_height = image.shape[:2]\n angle_radian = math.radians(angle)\n width = abs(np.sin(angle_radian) * old_height) + abs(np.cos(angle_radian) * old_width)\n height = abs(np.sin(angle_radian) * old_width) + abs(np.cos(angle_radian) * old_height)\n\n image_center = tuple(np.array(image.shape[1::-1]) / 2)\n rot_mat = cv2.getRotationMatrix2D(image_center, angle, 1.0)\n rot_mat[1, 2] += (width - old_width) / 2\n rot_mat[0, 2] += (height - old_height) / 2\n return cv2.warpAffine(image, rot_mat, (int(round(height)), int(round(width))), borderValue=background)\n\nimage = cv2.imread('input.png')\ngrayscale = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)\nangle = determine_skew(grayscale)\nrotated = rotate(image, angle, (0, 0, 0))\ncv2.imwrite('output.png', rotated)\n```\n\n## Debug images\n\nIf you get wrong skew angle you can generate debug images, that can help you to tune the skewing detection.\n\nIf you install deskew with `pip install deskew[debug_images]` you can get some debug images used for\nthe skew detection with the function `determine_skew_debug_images`.\n\nTo start the investigation you should first increase the `num_peaks` (default `20`) and use\nthe `determine_skew_debug_images` function.\n\nThen you can try to tune the following arguments `num_peaks`, `angle_pm_90`, `min_angle`, `max_angle`,\n`min_deviation` and eventually `sigma`.\n\nInspired by Alyn: https://github.com/kakul/Alyn\n\n## Contributing\n\nInstall the pre-commit hooks:\n\n```bash\npip install pre-commit\npre-commit install --allow-missing-config\n```\n\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "Skew detection and correction in images containing text",
"version": "1.5.1",
"project_urls": {
"Homepage": "https://github.com/sbrunner/deskew",
"Repository": "https://github.com/sbrunner/deskew"
},
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "4bb70dd282f6e83759939cd0ea78920cc67c7d57503f9ba2f51deaf05ecb1f20",
"md5": "903430a5f9bef400e2684d48de2e0c21",
"sha256": "4759427758c6a754077eed306c5ef22ebc9644d557748763f968cfd242ead18f"
},
"downloads": -1,
"filename": "deskew-1.5.1-py3-none-any.whl",
"has_sig": false,
"md5_digest": "903430a5f9bef400e2684d48de2e0c21",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.9",
"size": 7698,
"upload_time": "2024-01-10T16:37:30",
"upload_time_iso_8601": "2024-01-10T16:37:30.144103Z",
"url": "https://files.pythonhosted.org/packages/4b/b7/0dd282f6e83759939cd0ea78920cc67c7d57503f9ba2f51deaf05ecb1f20/deskew-1.5.1-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "2fdddc20f529fc1246e52e851a79ae718951134dda366c070e48c08c03654c3c",
"md5": "2b1b139555a930929e1cf60b32fa4a62",
"sha256": "b27d910da9ec34fa9abf867e3aed19dda87ab12529fc51124d38b7299ffaa387"
},
"downloads": -1,
"filename": "deskew-1.5.1.tar.gz",
"has_sig": false,
"md5_digest": "2b1b139555a930929e1cf60b32fa4a62",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.9",
"size": 7110,
"upload_time": "2024-01-10T16:37:32",
"upload_time_iso_8601": "2024-01-10T16:37:32.079907Z",
"url": "https://files.pythonhosted.org/packages/2f/dd/dc20f529fc1246e52e851a79ae718951134dda366c070e48c08c03654c3c/deskew-1.5.1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-01-10 16:37:32",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "sbrunner",
"github_project": "deskew",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "deskew"
}
JSON
