https://github.com/abdulfatir/pyblur
Soft Blurring in Python using OpenCV
https://github.com/abdulfatir/pyblur
Last synced: 5 months ago
JSON representation
Soft Blurring in Python using OpenCV
- Host: GitHub
- URL: https://github.com/abdulfatir/pyblur
- Owner: abdulfatir
- License: mit
- Created: 2016-06-15T09:11:55.000Z (over 9 years ago)
- Default Branch: master
- Last Pushed: 2016-06-17T15:04:51.000Z (over 9 years ago)
- Last Synced: 2025-03-31T11:41:21.682Z (6 months ago)
- Language: Python
- Homepage:
- Size: 1.98 MB
- Stars: 11
- Watchers: 2
- Forks: 5
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# pyBlur
`pyBlur` is a simple extension to OpenCV's `GaussianBlur`. There are times when you need to blur only a part of the image instead of blurring the entire image. `pyBlur` can blur any arbitrary shape in an image. You just need to pass the co-ordinates and it will blur the closed polygon formed by those co-ordinates. This comes in handly when you've to blur contours as you can just pass the result of `cv2.findContours` to `pyBlur`'s functions.
When a part of an image is blurred, most of the times it is harsh on the eyes. `pyBlur` also provides a method of blurring softly such that the blurred portion blends into the image in the background.
## Documentation
* `pyblur.**BlurRect**(image, rect, ksize, sigmaX [, sigmaY])`
Blurs a rectangular portion of the image defined by a numpy array `rect` which is `[x,y,width,height]`. `ksize`,`sigmaX`, and `sigmaY` mean the same as they mean in `cv2.GaussianBlur`.
* `pyblur.**BlurContours**(image, contours, ksize, sigmaX [, sigmaY])`
Blurs arbitrary polygons defined by the contours. `ksize`,`sigmaX`, and `sigmaY` mean the same as they mean in `cv2.GaussianBlur`.
* `pyblur.**SoftBlurRect**(image, rect, ksize, sigmaX [,sigmaY [,mksize [,msigmaX [,msigmaY]]]], **kwargs)`
Blurs a rectangular portion of the image defined by a numpy array `rect` which is `[x,y,width,height]` softly such that it blends with the background.`ksize`,`sigmaX`, and `sigmaY` are the parameters used to blur the image once using `cv2.GaussianBlur`.`mksize`,`msigmaX`, and `msigmaY` are the parameters used to blur the mask using `cv2.GaussianBlur` on successive iterations. `iters` can be passed to `kwargs` to set the number of iterations used for soft blurring.
* `pyblur.**SoftBlurContours**(image, contours, ksize, sigmaX [,sigmaY [,mksize [,msigmaX [,msigmaY]]]], **kwargs)`
Blurs a rectangular portion of the image defined by a numpy array `rect` which is `[x,y,width,height]` softly such that it blends with the background.`ksize`,`sigmaX`, and `sigmaY` are the parameters used to blur the image once using `cv2.GaussianBlur`.`mksize`,`msigmaX`, and `msigmaY` are the parameters used to blur the mask using `cv2.GaussianBlur` on successive iterations. `iters` can be passed to `kwargs` to set the number of iterations used for soft blurring.
## Usage
(Check samples in next section)
```python
out1 = pyblur.BlurRect(img, [500,50,100,150], 17, 5)
# Blurs rectangle with top-left point at (500,50) and of size (100,150).
# kernel = 17x17
# sigmaX = 5
```
```python
out2 = pyblur.BlurContours(img, [[np.array([[500,50],[600,300],[150,200]])]], 17, 5)
# Blurs contours passed as [[np.array([[500,50],[600,300],[150,200]])]].
# The array consists of one polygon formed by points (500,50), (600,300), and (150,200).
# kernel = 17x17
# sigmaX = 5
```
```python
out3 = pyblur.SoftBlurRect(img, [500,50,100,150], 27, 5)
# Soft Blurs rectangle with top-left point at (500,50) and of size (100,150).
# kernel = 27x27
# sigmaX = 5
```
```python
out4 = pyblur.SoftBlurContours(img, [[np.array([[500,50],[600,300],[150,200]])],
[np.array([[800,500],[950,500],[900,650]])]], 17, 5)
# Soft Blurs contours passed as [[np.array([[500,50],[600,300],[150,200]])]].
# The array consists of two polygons formed by points \
# [(500,50), (600,300), (150,200)] and [(800,500),(950,500),(900,650)].
# kernel = 17x17
# sigmaX = 5
```
```python
out5 = pyblur.SoftBlurRect(img, [500,50,100,150], 27, 5, 5, 55, iters=7)
# Soft Blurs rectangle with top-left point at (500,50) and of size (100,150).
# kernel for image = 27x27
# sigmaX for image = 5
# sigmaY for image = 5
# kernel for mask = 55x55
# no. of iterations = 7
```
## Results
Input
Output 1
Output 2
Output 3
Output 4
Output 5
