mpl-image-labeller
Use Matplotlib to label images for classification. Works anywhere Matplotlib does - from the notebook to a standalone gui!
For more see the documentation.
Install
pip install mpl-image-labeller
Key features
- Simple interface
- Uses keys instead of mouse
- Only depends on Matplotlib
- Works anywhere - from inside Jupyter to any supported GUI framework
- Displays images with correct aspect ratio
- Easily configurable keymap
- Smart interactions with default Matplotlib keymap
- Callback System (see
examples/callbacks.py)
single class per image
multiple classes per image
Usage
import matplotlib.pyplot as plt
import numpy as np
from mpl_image_labeller import image_labeller
images = np.random.randn(5, 10, 10)
labeller = image_labeller(
images, classes=["good", "bad", "meh"], label_keymap=["a", "s", "d"]
)
plt.show()
accessing the axis You can further modify the image (e.g. add masks over them) by using the plotting methods on axis object accessible by labeller.ax.
Lazy Loading Images If you want to lazy load your images you can provide a function to give the images. This function should take the integer idx as an argument and return the image that corresponds to that index. If you do this then you must also provide N_images in the constructor to let the object know how many images it should expect. See examples/lazy_loading.py for an example.
Controls
<-move one image back->move one image forward
To label images use the keys defined in the label_keymap argument - default 0, 1, 2...
Get the labels by accessing the labels property.
Overwriting default keymap
Matplotlib has default keybindings that it applied to all figures via rcparams.keymap that allow for actions such as s to save or q to quit. If you inlcude one of these keys as a shortcut for labelling as a class then that default keymap will be disabled for that figure.
Related Projects
This is not the first project to implement easy image labelling but seems to be the first to do so entirely in Matplotlib. The below projects implement varying degrees of complexity and/or additional features in different frameworks.
10 Jan 1, 2023
5 Jan 4, 2023
135 Dec 19, 2022
53 Nov 15, 2022
15 Dec 4, 2022
3 May 1, 2022
19 Dec 12, 2022
141 Dec 30, 2022
1k Dec 25, 2022
![[pre-commit.ci] pre-commit autoupdate](https://avatars.githubusercontent.com/in/68672?v=4)
51 Nov 28, 2022
53 Sep 17, 2022
1.2k Dec 29, 2022
73 Nov 30, 2022
0 Dec 07, 2022
73 Nov 06, 2022
1 Dec 02, 2021
4.9k Jan 05, 2023
57 Nov 24, 2022
110 Dec 20, 2022
37 Nov 10, 2022
137 Dec 31, 2022
314 Dec 30, 2022
8.7k Jan 08, 2023
27 Dec 16, 2022
13 Aug 18, 2022
71 Dec 30, 2022
166 Dec 28, 2022
9 Sep 01, 2022
2 Oct 20, 2022