def neural_network_test_table_gen():
X, y = default_img_set()
X = [[full_prepare(img) for img in same_sample] for same_sample in X]
Xs = count_blobs_with_all_methods(X)
Xs = [np.array(X_count) for X_count in Xs]
y = np.array(y)
row_names = ('Wszystkie detale', 'Śledzone detale',
'Stosunek śledzonych detali')
with open('exports/neural_network_test.csv', 'w') as csvfile:
filewriter = csv.writer(csvfile, delimiter=';')
# Header
filewriter.writerow(('Metoda zliczania detali', 'Wskaźnik',
'wskaźnik'))
filewriter.writerow(('Metoda zliczania detali', 'Błąd', 'Dokładność'))
for X, name in zip(Xs, row_names):
X = np.array(X)
X_train, X_test, y_train, y_test = train_test_split(
X, y, stratify=y, test_size=0.33, random_state=1)
model = default_grain_classifier_model()
model.compile(
optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
model.fit(X_train, y_train, epochs=300, verbose=0)
score = model.evaluate(X_test, y_test, verbose=0)
filewriter.writerow((name, *score))
def neural_network_validation_table_gen():
X, y = default_img_set()
X = [[full_prepare(img) for img in same_sample] for same_sample in X]
Xs = count_blobs_with_all_methods(X)
Xs = [np.array(X_count) for X_count in Xs]
y = np.array(y)
row_names = ('Wszystkie detale', 'Śledzone detale',
'Stosunek śledzonych detali')
with open('exports/neural_network_validation.csv', 'w') as csvfile:
filewriter = csv.writer(csvfile, delimiter=';')
# Header
filewriter.writerow(('Metoda zliczania detali', 'Wskaźnik',
'wskaźnik'))
filewriter.writerow(('Metoda zliczania detali', 'Błąd', 'Dokładność'))
for X, name in zip(Xs, row_names):
model = default_grain_classifier_model()
model.compile(
optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
scores = np.array(network_cross_validation(model, X, y, 3))
score = np.round(scores.mean(axis=0), 2)
filewriter.writerow((name, *score))
def neural_network_trainig_plots_gen():
X, y = default_img_set()
X = [[full_prepare(img) for img in same_sample] for same_sample in X]
Xs = count_blobs_with_all_methods(X)
files_suffixes = ('all', 'remaining', 'ratio')
for X, suffix in zip(Xs, files_suffixes):
X = np.array(X)
y = np.array(y)
X_train, _, y_train, _ = train_test_split(
X, y, stratify=y, test_size=0.33, random_state=1)
model = default_grain_classifier_model()
model.compile(
optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
history = model.fit(X_train, y_train, epochs=300, verbose=0)
plt.figure()
plt.title('Historia treningu modelu')
plt.xlabel('Epoka')
plt.plot(history.history['accuracy'], c='b')
plt.plot(history.history['loss'], c='r')
plt.legend(('Dokładność', 'Błąd'))
tikzplotlib.save('exports/neural_network_trainig_' + suffix)
def blob_analysis_plots_gen():
X, y = default_img_set()
X = [[full_prepare(img) for img in same_sample] for same_sample in X]
Xa, Xr, Xp = count_blobs_with_all_methods(X)
colors = ('r', 'g', 'b', 'y')
labels = ('E5R', 'E11R', 'E6R', 'E16R')
plot_blob_stat(Xa, y, colors)
plt.title('Liczba wszystkich detali')
plt.xlabel('Minuty')
plt.ylabel('Liczba detali')
patch_plot_legend(colors, labels)
tikzplotlib.save('exports/blob_analysis_all')
plot_blob_stat(Xr, y, colors)
plt.title('Liczba śledzonych detali')
plt.xlabel('Minuty')
plt.ylabel('Liczba detali')
patch_plot_legend(colors, labels)
tikzplotlib.save('exports/blob_analysis_remaining')
plot_blob_stat(Xp, y, colors)
plt.title('Pozostały procent śledzonych detali')
plt.xlabel('Minuty')
plt.ylabel('Liczba detali')
patch_plot_legend(colors, labels)
tikzplotlib.save('exports/blob_analysis_ratio')
def main():
'''Plot number of detected blobs using three ways of counting.'''
X, y = default_img_set()
X = [[full_prepare(img) for img in same_sample] for same_sample in X]
Xa, Xr, Xp = count_blobs_with_all_methods(X)
colors = ('r', 'g', 'b', 'y')
labels = ('E5R', 'E6R', 'E11R', 'E16R')
plot_blob_stat(Xa, y, colors)
plt.title('Number of all blobs')
plt.xlabel('minutes')
plt.ylabel('number of all detected blobs')
patch_plot_legend(colors, labels)
plot_blob_stat(Xr, y, colors)
plt.title('Number of remaining blobs')
plt.xlabel('minutes')
plt.ylabel('remaining blobs')
patch_plot_legend(colors, labels)
plot_blob_stat(Xp, y, colors)
plt.title('Ratio of remaining blobs')
plt.xlabel('minutes')
plt.ylabel('Ratio of remaining blobs')
patch_plot_legend(colors, labels)
plt.show()
def blob_ratio_table_gen():
sample_names = ('104_E5R', '106_E11R', '107_E6R', '111_E16R')
with open('exports/neural_network_comparison.csv', 'w') as csvfile:
filewriter = csv.writer(csvfile, delimiter=';')
# Header
filewriter.writerow(('Próbka', 'Minuta 0', 'Minuta 1', 'Minuta 2',
'Minuta 3', 'Minuta 4'))
for name in sample_names:
imgs = load_img_series('img/' + name)
imgs_prep = [full_prepare(img) for img in imgs]
stages_rem = find_blob_series(imgs_prep)
ratios = ratio_of_remaining_blobs_in_stages(stages_rem)
ratios = (round(ratio, 2) for ratio in ratios)
filewriter.writerow((name, *ratios))
def blob_detection_compare_plots_gen():
img = imread('img/104_E5R_0.jpg')
img_crop = crop_ui(rgb2gray(img))
img_prep = full_prepare(img)
blobs_list = compare_detection(img_prep)
suffixes = ('LoG', 'DoG', 'DoH')
for blobs, suffix in zip(blobs_list, suffixes):
_, ax = plt.subplots()
plt.title('Liczba wykrytych detali: {}'.format(len(blobs)))
plt.imshow(img_crop, cmap=plt.get_cmap('gray'))
for blob in blobs:
y, x, r = blob
c = plt.Circle((x, y), r, color='r', fill=False)
ax.add_patch(c)
ax.set_axis_off()
tikzplotlib.save('exports/blob_detection_compare_' + suffix)
def confusion_matrix_table_gen():
X, y = default_img_set()
X = [[full_prepare(img) for img in same_sample] for same_sample in X]
X = count_blobs_with_all_methods(X)[2]
X = np.array(X)
y = np.array(y)
model = default_grain_classifier_model()
model.compile(
optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
mcm = mean_confusion_matrix(model, X, y, 3)
np.savetxt(
"exports/mean_confusion_matrix_ratio.csv",
mcm,
fmt='%.2f',
delimiter=";")
def blob_count_plots_gen():
imgs = load_img_series('img/104_E5R')
imgs_prep = [full_prepare(img) for img in imgs]
imgs_crop = [crop_ui(rgb2gray(img)) for img in imgs]
stages_all = find_blob_series(imgs_prep, only_remaining=False)
stages_rem = find_blob_series(imgs_prep)
# Map stages on first image
colors = ('blue', 'blueviolet', 'magenta', 'crimson', 'red')
fig = plt.figure(frameon=False)
ax = fig.add_axes([0, 0, 1, 1])
plt.imshow(imgs_crop[0], cmap=plt.get_cmap('gray'))
for stage, color in zip(stages_rem, colors):
for blob in stage:
y, x, r = blob
c = plt.Circle((x, y), r, color=color, fill=False)
ax.add_patch(c)
ax.set_axis_off()
plt.savefig('exports/blob_tracker', dpi=300)
# Show two methods combined to compare
loop_set = enumerate(zip(stages_rem, stages_all, imgs_crop))
for i, (stage_rem, stage_all, img) in loop_set:
fig = plt.figure(frameon=False)
ax = fig.add_axes([0, 0, 1, 1])
plt.imshow(img, cmap=plt.get_cmap('gray'))
for blob_all in stage_all:
y, x, r = blob_all
c = plt.Circle((x, y), r, color='b', fill=False)
ax.add_patch(c)
for blob_rem in stage_rem:
y, x, r = blob_rem
c = plt.Circle((x, y), r, color='r', fill=False)
ax.add_patch(c)
ax.set_axis_off()
plt.savefig('exports/blob_tracker_min_' + str(i))
def network_comparison_table_gen():
X, y = default_img_set()
X = [[full_prepare(img) for img in same_sample] for same_sample in X]
X = [
ratio_of_remaining_blobs_in_stages(find_blob_series(img_series))
for img_series in X
]
X = np.array(X)
y = np.array(y)
with open('exports/neural_network_comparison.csv', 'w') as csvfile:
filewriter = csv.writer(csvfile, delimiter=';')
# Header
filewriter.writerow(('Parametr', 'Wartość', 'Błąd', 'Dokładność'))
# Activation functions
activation_funcs = ('sigmoid', 'relu', 'elu', 'tanh')
for func in activation_funcs:
model = keras.Sequential([
keras.layers.Dense(5, activation=func),
keras.layers.Dense(256, activation=func),
keras.layers.Dense(128, activation=func),
keras.layers.Dense(4, activation='softmax')
])
model.compile(
optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
scores = np.array(network_cross_validation(model, X, y, 3))
score = np.round(scores.mean(axis=0), 2)
filewriter.writerow(('Funkcja aktywacji', func, *score))
# Number of hidden layers
models = []
models.append(
keras.Sequential([
keras.layers.Dense(5, activation='tanh'),
keras.layers.Dense(512, activation='tanh'),
keras.layers.Dense(4, activation='softmax')
]))
models.append(
keras.Sequential([
keras.layers.Dense(5, activation='tanh'),
keras.layers.Dense(256, activation='tanh'),
keras.layers.Dense(128, activation='tanh'),
keras.layers.Dense(4, activation='softmax')
]))
for model, i in zip(models, (1, 2)):
model.compile(
optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
scores = np.array(network_cross_validation(model, X, y, 3))
score = np.round(scores.mean(axis=0), 2)
filewriter.writerow(('Liczba warstw ukrytych', i, *score))
# Number of neurons in hidden layers
neurons_num = ((128, 64), (256, 128), (512, 126))
for num in neurons_num:
model = keras.Sequential([
keras.layers.Dense(5, activation='tanh'),
keras.layers.Dense(num[0], activation='tanh'),
keras.layers.Dense(num[1], activation='tanh'),
keras.layers.Dense(4, activation='softmax')
])
model.compile(
optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
scores = np.array(network_cross_validation(model, X, y, 3))
score = np.round(scores.mean(axis=0), 2)
filewriter.writerow(('Liczba neuronów w warstwach ukrytych',
'{} i {}'.format(num[0], num[1]), *score))
# Optimizer
model = default_grain_classifier_model()
optimizers = ('sgd', 'adam')
for opt in optimizers:
model.compile(
optimizer=opt,
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
scores = np.array(network_cross_validation(model, X, y, 3))
score = np.round(scores.mean(axis=0), 2)
filewriter.writerow(('Algorytm uczenia', opt, *score))
def main():
'''Demo blob tracking with various ways of counting blobs.'''
# Load images
imgs = load_img_series('img/104_E5R')
# Prepare images for processing
imgs_prep = [full_prepare(img) for img in imgs]
# Prepare cropped images for displaying
imgs_crop = [crop_ui(rgb2gray(img)) for img in imgs]
# Find blobs for stages of cooling with preserving only remainig ones
stages_rem = find_blob_series(imgs_prep)
# Map stages on first image
colors = ('blue', 'blueviolet', 'magenta', 'crimson', 'red')
_, ax = plt.subplots(1)
plt.title("Blobs detection with DoH")
plt.imshow(imgs_crop[0], cmap=plt.get_cmap('gray'))
for stage, color in zip(stages_rem, colors):
for blob in stage:
y, x, r = blob
c = plt.Circle((x, y), r, color=color, linewidth=0.75, fill=False)
ax.add_patch(c)
labels = ('Minute 0', 'Minute 1', 'Minute 2', 'Minute 3', 'Minute 4')
patch_plot_legend_outside(colors, labels)
print(ratio_of_remaining_blobs_in_stages(stages_rem))
# Show stages on subplots
_, ax = plt.subplots(2, 3, figsize=(12, 7))
ax = ax.flatten()
for idx, (stage, img) in enumerate(zip(stages_rem, imgs_crop)):
ax[idx].imshow(img, cmap=plt.get_cmap('gray'))
ax[idx].set_title("Minute: {}, blobs: {}".format(idx, len(stage)))
for blob in stage:
y, x, r = blob
c = plt.Circle((x, y), r, color='r', linewidth=0.75, fill=False)
ax[idx].add_patch(c)
ax[-1].set_axis_off()
plt.tight_layout()
# Find all blobs for every stage of cooling
stages_all = find_blob_series(imgs_prep, only_remaining=False)
# Show stages on subplots
_, ax = plt.subplots(2, 3, figsize=(12, 7))
ax = ax.flatten()
for idx, (stage, img) in enumerate(zip(stages_all, imgs_crop)):
ax[idx].imshow(img, cmap=plt.get_cmap('gray'))
ax[idx].set_title("Minute: {}, blobs: {}".format(idx, len(stage)))
for blob in stage:
y, x, r = blob
c = plt.Circle((x, y), r, color='r', linewidth=0.75, fill=False)
ax[idx].add_patch(c)
ax[-1].set_axis_off()
plt.tight_layout()
# Show two methods combined to compare
_, ax = plt.subplots(2, 3, figsize=(10, 7))
ax = ax.flatten()
# Show stages on subplots
loop_set = enumerate(zip(stages_rem, stages_all, imgs_crop))
for idx, (stage_rem, stage_all, img) in loop_set:
ax[idx].imshow(img, cmap=plt.get_cmap('gray'))
ax[idx].set_title("Minute: {}, all blobs: {}, rem blobs: {}".format(
idx, len(stage_all), len(stage_rem)))
for blob_all in stage_all:
y, x, r = blob_all
c = plt.Circle((x, y), r, color='b', linewidth=0.75, fill=False)
ax[idx].add_patch(c)
for blob_rem in stage_rem:
y, x, r = blob_rem
c = plt.Circle((x, y), r, color='r', linewidth=0.75, fill=False)
ax[idx].add_patch(c)
ax[-1].set_axis_off()
plt.tight_layout()
plt.show()
