def character_classification():
print('Loading data...')
x, y = load_data_chars()
print('Processing data..')
print('Training data shape: ', x.shape)
print('Test data shape: ', y.shape)
plots.plot_filters(x[0])
SVM.svm(x, y)
Naive_Bayes.naive_bayes(x, y)
KNN.knn(x, y)
CNN.fit_cnn(x, y, trials=1, network_type='simple')
ax[0, 0].axis('off')
ax[0, 0].set_title('Image', fontsize=10)
ax[0, 1].imshow(add_edge_detection_filter(img), cmap='gray')
ax[0, 1].axis('off')
ax[0, 1].set_title('Edge detection filter', fontsize=10)
ax[0, 2].imshow(add_otsu_filter(img), cmap='gray')
ax[0, 2].axis('off')
ax[0, 2].set_title('Otsu filter', fontsize=10)
ax[1, 0].imshow(add_gaussian_filter(img), cmap='gray')
ax[1, 0].axis('off')
ax[1, 0].set_title('Gaussian filter', fontsize=10)
ax[1, 1].imshow(add_bilateral_filter(img), cmap='gray')
ax[1, 1].axis('off')
ax[1, 1].set_title('Bilateral filter', fontsize=10)
ax[1, 2].imshow(add_median_filter(img), cmap='gray')
ax[1, 2].axis('off')
ax[1, 2].set_title('Median filter', fontsize=10)
fig.tight_layout()
plt.show()
if __name__ == '__main__':
x, y = load_data_chars()
plot_filters(x[2])
print('=== Convolution Neural Network ===')
test_accuracy = np.zeros(trials)
running_time = np.zeros(trials)
x = scale_input(x)
x = grey_scale(x)
x = add_dimension(x)
x, y = add_pictures_without_chars(x, y)
y = to_categorical(y, int(np.max(y) + 1))
for i in range(trials):
print('Training network ', i + 1)
start = time.time()
random_state = 100 + i
x_train, x_test, y_train, y_test = train_test_split(
x, y, test_size=0.2, random_state=random_state, stratify=y)
network = CNN(num_classes=len(y[0]),
sample=x_train[0],
network_type=network_type)
network.train_generator(x_train, y_train, model_weights)
test_accuracy[i] = network.test(x_test, y_test, model_weights)
running_time[i] = time.time() - start
print('Running time: ', running_time[i])
print('Average test accuracy over ', trials, ' trials: ',
np.mean(test_accuracy))
print('Average running time over ', trials, ' trials: ',
np.mean(running_time))
if __name__ == "__main__":
img, target = load_data_chars()
fit_cnn(img, target, trials=1)