def main(argv):
print(tf.version.VERSION)
image_size = 224
test_run = 'zC'
# load the data
(x_train, y_train), (x_test, y_test) = odir.load_data(image_size, 1)
class_names = [
'Normal', 'Diabetes', 'Glaucoma', 'Cataract', 'AMD', 'Hypertension',
'Myopia', 'Others'
]
# plot data input
plotter = Plotter(class_names)
plotter.plot_input_images(x_train, y_train)
x_test_drawing = x_test
# normalize input based on model
normalizer = Normalizer()
x_test = normalizer.normalize_vgg16(x_test)
# load one of the test runs
model = tf.keras.models.load_model(
r'C:\Users\thund\Source\Repos\TFM-ODIR\models\image_classification\modelvgg100.h5'
)
model.summary()
# display the content of the model
baseline_results = model.evaluate(x_test, y_test, verbose=2)
for name, value in zip(model.metrics_names, baseline_results):
print(name, ': ', value)
print()
# test a prediction
test_predictions_baseline = model.predict(x_test)
plotter.plot_confusion_matrix_generic(y_test, test_predictions_baseline,
test_run, 0)
# save the predictions
prediction_writer = Prediction(test_predictions_baseline, 400)
prediction_writer.save()
prediction_writer.save_all(y_test)
# show the final score
score = FinalScore()
score.output()
# plot output results
plotter.plot_output(test_predictions_baseline, y_test, x_test_drawing)
plt.plot(history.history['accuracy'], label='accuracy')
plt.plot(history.history['val_accuracy'], label='val_accuracy')
plt.xlabel('Epoch')
plt.ylabel('Accuracy')
plt.legend(loc='lower right')
plt.savefig(os.path.join(newfolder, 'plot2.png'))
plt.show()
# display the content of the model
baseline_results = model.evaluate(x_test, y_test, verbose=2)
for name, value in zip(model.metrics_names, baseline_results):
print(name, ': ', value)
print()
# test a prediction
test_predictions_baseline = model.predict(x_test)
plotter.plot_confusion_matrix_generic(y_test, test_predictions_baseline,
os.path.join(newfolder, 'plot3.png'), 0)
# save the predictions
prediction_writer = Prediction(test_predictions_baseline, 400, newfolder)
prediction_writer.save()
prediction_writer.save_all(y_test)
# show the final score
score = FinalScore(newfolder)
score.output()
# plot output results
plotter.plot_output(test_predictions_baseline, y_test, x_test_drawing,
os.path.join(newfolder, 'plot4.png'))
def main(argv):
print(tf.version.VERSION)
image_size = 128
(x_train, y_train), (x_test, y_test) = odir.load_data(image_size)
x_train, x_test = x_train / 255.0, x_test / 255.0
x_train = (x_train - x_train.mean()) / x_train.std()
x_test = (x_test - x_test.mean()) / x_test.std()
defined_metrics = [
tf.keras.metrics.BinaryAccuracy(name='accuracy'),
tf.keras.metrics.Precision(name='precision'),
tf.keras.metrics.Recall(name='recall'),
tf.keras.metrics.AUC(name='auc'),
]
factory = Factory((image_size, image_size, 3), defined_metrics)
model = factory.compile(ModelTypes.inception_v1)
class_names = [
'Normal', 'Diabetes', 'Glaucoma', 'Cataract', 'AMD', 'Hypertension',
'Myopia', 'Others'
]
# plot data input
plotter = Plotter(class_names)
print("Training")
class_weight = {
0: 1.,
1: 1.583802025,
2: 8.996805112,
3: 10.24,
4: 10.05714286,
5: 14.66666667,
6: 10.7480916,
7: 2.505338078
}
callback = tf.keras.callbacks.EarlyStopping(monitor='val_loss',
patience=3,
mode='min',
verbose=1)
history = model.fit_generator(generator=generator(x_train, y_train),
steps_per_epoch=len(x_train),
epochs=30,
verbose=1,
callbacks=[callback],
validation_data=(x_test, y_test),
shuffle=True)
print("plotting")
plotter.plot_metrics(history, 'inception_1', 2)
print("saving")
model.save('model_inception_30.h5')
# Hide meanwhile for now
plt.plot(history.history['accuracy'], label='accuracy')
plt.plot(history.history['val_accuracy'], label='val_accuracy')
plt.xlabel('Epoch')
plt.ylabel('Accuracy')
plt.legend(loc='lower right')
plt.savefig('image_run2' + 'inception_1' + '.png')
plt.show()
test_loss, test_acc = model.evaluate(x_test, y_test, verbose=2)
print(test_acc)
test_predictions_baseline = model.predict(x_test)
plotter.plot_confusion_matrix_generic(y_test, test_predictions_baseline,
'inception_1', 0)
# save the predictions
prediction_writer = Prediction(test_predictions_baseline, 400)
prediction_writer.save()
prediction_writer.save_all(y_test)
# show the final score
score = FinalScore()
score.output()