K-SVD

An K-SVD implementaion written in Python.

Build Status PyPI

Installation

pip install ksvd

Usage

import numpy as np
from ksvd import ApproximateKSVD

# X ~ gamma.dot(dictionary)
X = np.random.randn(1000, 20)
aksvd = ApproximateKSVD(n_components=128)
dictionary = aksvd.fit(X).components_
gamma = aksvd.transform(X)

Feature

Example

License

Licensed under the Apache License 2.0.

References