print "Printing digit example #" + str(example_number + 1) + " with label: " \
+ str(ALL_TRAINING_LABELS[example_number])
_print_digit_image(ALL_TRAINING_IMAGES[example_number])
print
for i in range(10, 11):
# Calling the basic feature extractor
features = mp.extract_basic_features(
ALL_TRAINING_IMAGES[example_number], DATA_WIDTH, DATA_HEIGHT)
# Compute parameters for a Naive Bayes classifier using the basic feature
# extractor
mp.compute_statistics(ALL_TRAINING_IMAGES,
ALL_TRAINING_LABELS,
DATA_WIDTH,
DATA_HEIGHT,
mp.extract_advanced_features,
percentage=10.0 * i,
k=.0001)
# Making predictions on validation data
predicted_labels = mp.classify(ALL_VALIDATION_IMAGES, DATA_WIDTH,
DATA_HEIGHT,
mp.extract_advanced_features)
incorrect = []
correct_count = 0.0
for ei in range(len(predicted_labels)):
if (ALL_VALIDATION_LABELS[ei] == predicted_labels[ei]):
correct_count += 1
else:
# Calling the basic feature extractor
features = mp.extract_basic_features(ALL_TRAINING_IMAGES[example_number],
DATA_WIDTH, DATA_HEIGHT)
#mp.extract_advanced_features(ALL_TRAINING_IMAGES[example_number],DATA_WIDTH, DATA_HEIGHT)
#mp.extract_advanced_features(ALL_TRAINING_IMAGES[2075],DATA_WIDTH, DATA_HEIGHT)
# Compute parameters for a Naive Bayes classifier using the basic feature
# extractor
#mp.compute_statistics(ALL_TRAINING_IMAGES, ALL_TRAINING_LABELS, DATA_WIDTH,DATA_HEIGHT, mp.extract_basic_features)
# Making predictions on validation data
#predicted_labels = mp.classify(ALL_VALIDATION_IMAGES, DATA_WIDTH, DATA_HEIGHT,mp.extract_basic_features)
#with advance features
#mp.compute_statistics(ALL_TRAINING_IMAGES, ALL_TRAINING_LABELS, DATA_WIDTH, DATA_HEIGHT, mp.extract_advanced_features)
#predicted_labels = mp.classify(ALL_VALIDATION_IMAGES, DATA_WIDTH, DATA_HEIGHT, mp.extract_advanced_features)
#with final features
mp.compute_statistics(ALL_TRAINING_IMAGES, ALL_TRAINING_LABELS, DATA_WIDTH,
DATA_HEIGHT, mp.extract_final_features)
predicted_labels = mp.classify(ALL_VALIDATION_IMAGES, DATA_WIDTH,
DATA_HEIGHT, mp.extract_final_features)
correct_count = 0.0
for ei in range(len(predicted_labels)):
if (ALL_VALIDATION_LABELS[ei] == predicted_labels[ei]):
correct_count += 1
print "Correct prediction: " + str(correct_count / len(predicted_labels))
print("--- %s seconds ---" % (time.time() - start_time))
# Load all data
_load_all_data()
# Pring a random traning example
example_number = random.randint(0, NUMBER_OF_TRAINING_EXAMPLES)
print "Printing digit example #" + str(example_number + 1) + " with label: " \
+ str(ALL_TRAINING_LABELS[example_number])
_print_digit_image(ALL_TRAINING_IMAGES[example_number])
print
# Calling the basic feature extractor
features = mp.extract_basic_features(ALL_TRAINING_IMAGES[example_number],
DATA_WIDTH, DATA_HEIGHT)
# Compute parameters for a Naive Bayes classifier using the basic feature
# extractor
mp.compute_statistics(ALL_TRAINING_IMAGES, ALL_TRAINING_LABELS, DATA_WIDTH,
DATA_HEIGHT, mp.extract_basic_features, 100)
# Making predictions on validation data
predicted_labels = mp.classify(ALL_VALIDATION_IMAGES, DATA_WIDTH,
DATA_HEIGHT, mp.extract_basic_features)
correct_count = 0.0
for ei in range(len(predicted_labels)):
if (ALL_VALIDATION_LABELS[ei] == predicted_labels[ei]):
correct_count += 1
print "Correct prediction: " + str(correct_count / len(predicted_labels))