quantum-image-classifier 0.1.0

Quatum image classifier: A library of different quantum algorithms used to classify images

Quantum image classifier

Data use

You can generate synthetic data by calling the function generate_synthetic_data(n_dim: int, n_clusters: int, m_samples: int) implemented in data_generator.py. You have to be aware that, in order to Nearest Centroid to work, n_dim has to be power of 2. This function returns a set of m_samples vectors X with a set of labels y associated with the vector in the same possition on X. Example:

X, y = generate_synthetic_data(8, 4, 250)
train_X = X[:200]
train_y = y[:200]
test_X = X[200:]
test_y = y[200:]

If you want, you can also use the MNIST dataset with a PCA function used to reduce the dimension to n components calling get_MNIST(n_components) implemented in data_loader.py. Same as with the synthetic data, you have to be aware to use only an power of 2 to make Nearest Centroid work. Example:

train_X, train_y, test_X, test_y = get_MNIST(8)

Classifiers

Nearest centroid

Once you get the data, you need to create the object NearestCentroid with the training dataset that you want. After that, you can call the function predict(self, X: np.ndarray) owned by the defined object. Example:

train_X, train_y, test_X, test_y = get_MNIST(8)
nearest_centroid = NearestCentroid(train_X, train_y, n_dim)
labels_predicted = nearest_centroid.predict(test_X)