https://github.com/freakwill/image-encoding
image encoding by dimensional reduction models (in scikit-learn)
https://github.com/freakwill/image-encoding
Last synced: about 1 month ago
JSON representation
image encoding by dimensional reduction models (in scikit-learn)
- Host: GitHub
- URL: https://github.com/freakwill/image-encoding
- Owner: Freakwill
- License: mit
- Created: 2024-04-26T05:06:24.000Z (about 2 years ago)
- Default Branch: main
- Last Pushed: 2024-08-11T11:01:03.000Z (almost 2 years ago)
- Last Synced: 2025-03-21T17:55:47.906Z (over 1 year ago)
- Language: Python
- Size: 34.2 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# image-encoding
image encoding by dimensional reduction models (in scikit-learn)
## Examples
### 1
```python
def demo_digit(folder=pathlib.Path.cwd(), *args, **kwargs):
if isinstance(folder, str):
folder = pathlib.Path(folder)
from sklearn import datasets
digists = datasets.load_digits()
model = FastICA(*args, **kwargs) # the backend model
enc = ImageEncoder(model, size=(8,8))
ip.fit(digists.data * (255//16))
save_in(enc.eigen_images, folder / 'eigen', exist_ok=True)
# generate new images
save_in(enc.generate(10, toimage=True), folder / 'eigen', exist_ok=True, prefix='generated')
demo_digit(n_components=15)
```
### 2
```python
def demo_face(folder=pathlib.Path.cwd(), *args, **kwargs):
# save images in the folder before demo
if isinstance(folder, str):
folder = pathlib.Path(folder)
# define a backend model, such as PCA, NMF
model = PCA(*args, **kwargs)
enc = ImageEncoder(model)
# a user-friendly API calling `fit` method of the model
enc.ezfit(folder=folder) # folder where the images are stored
# save the eigen images in `eigen/` subfolder
save_in(enc.eigen_images, folder / 'eigen', exist_ok=True)
# generate new images and save them in `generated/` subfolder
save_in(enc.generate(10, toimage=True), folder / 'generated', exist_ok=True)
# save the images (with the same size and mode) in the current path or a special folder
demo_face(n_components=10)
```