def _train_model(self, model_file=None):
print("Creating GALA feature manager...")
fm = features.moments.Manager()
fh = features.histogram.Manager(25, 0, 1, [0.1, 0.5, 0.9]) # Recommended numbers in the repo
fg = features.graph.Manager()
fc = features.contact.Manager()
self.fm = features.base.Composite(children=[fm, fh, fg, fc])
if model_file is not None and os.path.isfile(model_file):
print('Loading model from path ...')
rf = classify.load_classifier(model_file)
else:
gt, pr, sv = (map(imio.read_h5_stack, [self.gt, self.mem, self.sp]))
print("Creating training RAG...")
g_train = agglo.Rag(sv, pr, feature_manager=self.fm)
print("Learning agglomeration...")
(X, y, w, merges) = g_train.learn_agglomerate(gt, self.fm, learning_mode='permissive',
min_num_epochs=self.min_ep)[0]
y = y[:, 0]
rf = classify.DefaultRandomForest().fit(X, y)
# Save if path requested
if model_file is not None:
classify.save_classifier(rf, model_file)
self.model = agglo.classifier_probability(self.fm, rf)
gt_test, pr_test, p4_test, ws_test = map(imio.read_h5_stack, ['example-data/test-gt.lzf.h5', 'example-data/test-p1.lzf.h5', 'example-data/test-p4.lzf.h5', 'example-data/test-ws.lzf.h5'])
fm = features.moments.Manager()
fh = features.histogram.Manager()
fc = features.base.Composite(children=[fm, fh])
g_train = agglo.Rag(ws_train, pr_train, feature_manager=fc)
np.random.RandomState(0)
(X, y, w, merges) = map(np.copy, map(np.ascontiguousarray,
g_train.learn_agglomerate(gt_train, fc)[0]))
print X.shape
np.savez('example-data/train-set.npz', X=X, y=y)
y = y[:, 0]
rf = classify.DefaultRandomForest()
X.shape
np.random.RandomState(0)
rf = rf.fit(X, y)
classify.save_classifier(rf, 'example-data/rf-1.joblib')
learned_policy = agglo.classifier_probability(fc, rf)
g_test = agglo.Rag(ws_test, pr_test, learned_policy, feature_manager=fc)
g_test.agglomerate(0.5)
seg_test1 = g_test.get_segmentation()
imio.write_h5_stack(seg_test1, 'example-data/test-seg1.lzf.h5', compression='lzf')
g_train4 = agglo.Rag(ws_train, p4_train, feature_manager=fc)
np.random.RandomState(0)
(X4, y4, w4, merges4) = map(np.copy, map(np.ascontiguousarray,
g_train4.learn_agglomerate(gt_train, fc)[0]))
print X4.shape
np.savez('example-data/train-set4.npz', X=X4, y=y4)
y4 = y4[:, 0]
rf4 = classify.DefaultRandomForest()
np.random.RandomState(0)
rf4 = rf4.fit(X4, y4)
def train_and_save_classifier(training_data_file, filename,
classifier_kind='random forest'):
X, y = load_training_data(training_data_file)
cl = classify.get_classifier(classifier_kind)
cl.fit(X, y)
classify.save_classifier(cl, filename, use_joblib=False)
def train_and_save_classifier(training_data_file, filename,
classifier_kind='random forest'):
X, y = load_training_data(training_data_file)
cl = classify.get_classifier(classifier_kind)
cl.fit(X, y)
classify.save_classifier(cl, filename, use_joblib=False)
im = scipy.io.loadmat(inFileImage)['im']
im = im.astype('int32')
membraneTrain = scipy.io.loadmat(inFileMembrane)['membrane']
membraneTrain = membraneTrain.astype('float32')
gt_train = scipy.io.loadmat(inFileTruth)['truth']
gt_train = gt_train.astype('int64') #just in case!
ws_train = scipy.io.loadmat(inFileWatershed)['ws']
ws_train = ws_train.astype('int64') #just in case!
xdim, ydim, zdim = (im.shape)
fc = features.base.Composite(children=[features.moments.Manager(), features.histogram.Manager(25, 0, 1, [0.1, 0.5, 0.9]),
features.graph.Manager(), features.contact.Manager([0.1, 0.5, 0.9]) ])
print "Creating RAG..."
# create graph and obtain a training dataset
g_train = agglo.Rag(ws_train, membraneTrain, feature_manager=fc)
print 'Learning agglomeration...'
(X, y, w, merges) = g_train.learn_agglomerate(gt_train, fc,min_num_epochs=5)[0]
y = y[:, 0] # gala has 3 truth labeling schemes, pick the first one
print(X.shape, y.shape) # standard scikit-learn input format
print "Training classifier..."
# train a classifier, scikit-learn syntax
rf = classify.DefaultRandomForest().fit(X, y)
# a policy is the composition of a feature map and a classifier
learned_policy = agglo.classifier_probability(fc, rf)
classify.save_classifier(rf,'/mnt/pipeline/tools/i2g/packages/gala/ac4_full_classifier_v2.rf')