from tensorflow.python.keras.utils import plot_model
plot_model(model, to_file='keras-logos-gen-model.png', show_shapes=True, show_layer_names=True)
model.summary()
model.fit_generator(train_dir_iterator, epochs=20)
#print(model.evaluate_generator(test_dir_iterator, verbose=1))
with open("keras-logos-gen-predictions.txt", 'w') as out:
idx = 0
for batch in test_dir_iterator:
filenames = test_dir_iterator.filenames[idx : idx + test_dir_iterator.batch_size]
if len(filenames) == 0:
break
idx += test_dir_iterator.batch_size
predictions = model.predict_on_batch(batch[0])
for i in range(len(predictions)):
truth = test_class_map[np.argmax(batch[1][i])]
conf = np.max(predictions[i])
if conf > 0.25:
predicted = train_class_map[np.argmax(predictions[i])]
else:
predicted = "no-logo"
m = re.match(r".*/(\d+)\.jpg", filenames[i])
fid = m.group(1)
print("file = %s\ttruth = %s\tpredicted = %s\tconfidence = %f" % (filenames[i], truth, predicted, conf))
out.write("%s, %s, %f\n" % (fid, predicted, conf))
plt.plot(history.history['val_mean_absolute_error'],
label='MSE (validation data)')
plt.title('MAE for celab2000')
plt.ylabel('MAE value')
plt.xlabel('No. epoch')
plt.legend(loc="upper left")
plt.show()
#predictions
#getting the input as list
inputs = genfromtxt('cielab2000_sample_input-15000.csv', delimiter=',')
start1 = time.time() #pin starting point
#model prediction
prediction = model.predict_on_batch(np.array(inputs))
end1 = time.time() #pin ending point
print(prediction)
print(end1 - start1)
# print(end1-start1)
# print(results1)
results2 = list()
#color math delta calculation
start2 = time.time() #pin starting point
for j in range(0, len(inputs)):
l1, a1, b1, l2, a2, b2 = np.array(inputs[j])
output = getdelta_e_cie2000(l1, a1, b1, l2, a2, b2)
rounded_output = np.round(output, 5)
results2.append(rounded_output)
norm = [float(k) / max(results2) for k in results2]
