def convert_leaves_all_probs_pred(image, leaves, all_probs, num_leaves):
import imseg
preds = [] # NOTE(brandyn): This has the ILP turned off
preds = np.ascontiguousarray(preds, dtype=np.float64)
out0 = imseg.convert_labels_to_integrals(leaves, num_leaves)
out1 = imseg.convert_all_probs_to_integrals(all_probs)
return preds, np.ascontiguousarray(np.dstack([out0, out1]))
def convert_leaves_all_probs_pred_old(image, leaves, all_probs, num_leaves, classifiers_fn=None):
global CLASSIFIERS, CLASSIFIER_FEATURE
if classifiers_fn is None:
classifiers_fn = os.environ['CLASSIFIERS_FN']
get_classifier_confidence = lambda x: x[0][0] * x[0][1]
if CLASSIFIERS is None:
all_classifiers = sorted(file_parse.load(classifiers_fn))
name_classifiers = []
for x in range(len(all_classifiers)):
if x < len(all_classifiers): # TODO(brandyn): Fix memory issue so that we can use the last classifier too
name_classifiers.append((all_classifiers[x][0],
classifiers.loads(all_classifiers[x][1])))
else:
name_classifiers.append((all_classifiers[x][0],
name_classifiers[-1][1]))
all_classifiers[x] = None # NOTE(Brandyn): This is done to save memory
print('ILP Classifiers %r' % ([x for x, _ in name_classifiers],))
CLASSIFIERS = [x for _, x in name_classifiers]
if CLASSIFIER_FEATURE is None:
CLASSIFIER_FEATURE = features.select_feature('bovw_hog')
feature = CLASSIFIER_FEATURE(np.ascontiguousarray(image[:, :, :3]))
preds = np.ascontiguousarray([get_classifier_confidence(classifier.predict(feature))
for classifier in CLASSIFIERS], dtype=np.float64)
out0 = imseg.convert_labels_to_integrals(leaves, num_leaves)
out1 = imseg.convert_all_probs_to_integrals(all_probs)
return preds, np.ascontiguousarray(np.dstack([out0, out1]))
def convert_leaves_all_probs_pred(image, leaves, all_probs, num_leaves, classifiers_fn=None):
global CLASSIFIER_FEATURE, ALL_CLASSIFIERS, CLASSIFIERS
preds = []
if classifiers_fn:
get_classifier_confidence = lambda x: x[0][0] * x[0][1]
if ALL_CLASSIFIERS is None:
ALL_CLASSIFIERS = sorted(file_parse.load(classifiers_fn))
if CLASSIFIER_FEATURE is None:
CLASSIFIER_FEATURE = features.select_feature('bovw_hog')
feature = CLASSIFIER_FEATURE(np.ascontiguousarray(image[:, :, :3]))
for x in range(len(ALL_CLASSIFIERS)):
try:
classifier = CLASSIFIERS[x]
except KeyError:
classifier = classifiers.loads(ALL_CLASSIFIERS[x][1])
if x < 14:
CLASSIFIERS[x] = classifier
preds.append(get_classifier_confidence(classifier.predict(feature)))
preds = np.ascontiguousarray(preds, dtype=np.float64)
out0 = imseg.convert_labels_to_integrals(leaves, num_leaves)
out1 = imseg.convert_all_probs_to_integrals(all_probs)
return preds, np.ascontiguousarray(np.dstack([out0, out1]))
def map(self, k, v):
"""
Args:
k: Filename
v: (image_data, label_points)
Yields:
(k, (integral, label_points))
"""
image_data, label_points = v
argmax, max_probs, leaves, all_probs = self.tp.predict(image_data, all_probs=True, leaves=True)
print("Leaves[%s] AllProbs[%s]" % (str(leaves.shape), str(all_probs.shape)))
if self.integral_type == "argmax":
out = np.ascontiguousarray(imseg.convert_labels_to_integrals(argmax, self.tp.num_classes))
elif self.integral_type == "argmax_prob":
out = np.ascontiguousarray(imseg.convert_labels_probs_to_integrals(argmax, max_probs, self.tp.num_classes))
elif self.integral_type == "all_prob":
out = np.ascontiguousarray(imseg.convert_all_probs_to_integrals(all_probs))
elif self.integral_type == "spatial":
out = convert_leaves_all_probs_pred(image_data, leaves, all_probs, self.tp.num_leaves)
yield k, (out, label_points)
