https://github.com/ducha-aiki/extract-patches-old

Simple function for local patch extraction from OpenCV keypoints.
https://github.com/ducha-aiki/extract-patches-old

local-features opencv-python python sift

Last synced: 8 months ago
JSON representation

Simple function for local patch extraction from OpenCV keypoints.

• Host: GitHub
• URL: https://github.com/ducha-aiki/extract-patches-old
• Owner: ducha-aiki
• License: mit
• Created: 2019-06-21T12:50:16.000Z (almost 7 years ago)
• Default Branch: master
• Last Pushed: 2019-07-10T16:39:49.000Z (almost 7 years ago)
• Last Synced: 2025-05-06T23:31:24.597Z (about 1 year ago)
• Topics: local-features, opencv-python, python, sift
• Language: Jupyter Notebook
• Size: 8.55 MB
• Stars: 14
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# extract-patches

Simple function for local patch extraction from local features keypoints.

Accepts following formats:

- OpenCV keypoints

- Ellipse format [x y a b c]

- Affine features format [x y a11 a12 a21 a22]

- OpenCV keypoints + A (a11 a12 a21, a22), say from AffNet output

Usage:

    

```python

from extract_patches import extract_patches

PATCH_SIZE = 32

mrSize = 3.0

patches = extract_patches(kps1, img1, PATCH_SIZE, mrSize, 'cv2')

```

![extracted patches example](img/example.png)

See another example in this [notebook](patch-extraction-demo.ipynb)

Thanks to [Vassileios](https://github.com/vbalnt) for the baseline implementation.