def create_forest():
""" Create new random decision forest
"""
print('Creating random forest')
images = read_images()
# assign classes
image_classes = [icls for icls, _ in images]
# extract features
feature_vectors = []
for _, img_path in images:
cur = get_feature_vector(img_path)
feature_vectors.append(cur)
assert len(feature_vectors) > 0, 'Must have at least one feature vector to train on'
# good estimator number for classification tasks
sqrt_feat_num = int(np.sqrt(len(feature_vectors[0])))
print(' ', sqrt_feat_num, 'estimator%s' % ('' if sqrt_feat_num == 1 else 's'))
# create forest
clf = RandomForestClassifier(
n_estimators=sqrt_feat_num, n_jobs=-1,
max_depth=None, min_samples_split=1
)
clf = clf.fit(feature_vectors, image_classes)
return clf
def process_batch_custom(datum, max_people):
features = datum['features']
labels = datum['labels']
batch_x = []
batch_y = []
batch_pad = []
for member in range(labels.shape[0]):
pair_feat = []
pair_membership = []
pair_pad = []
for agent in range(labels.shape[0]):
pair_feat.append(
get_feature_vector(features[member], features[agent]))
pair_membership.append(labels[member] == labels[agent])
if member == agent:
pair_pad.append(-2)
else:
pair_pad.append(0)
n_remaining = max_people - len(pair_membership)
for dummy in range(n_remaining):
pair_feat.append(pair_feat[0])
pair_membership.append(pair_membership[0])
pair_pad.append(-2)
batch_x.append(np.array(pair_feat).astype(np.float32))
batch_y.append(
np.expand_dims(np.array(pair_membership).astype(np.float32),
axis=1))
batch_pad.append(
np.expand_dims(np.array(pair_pad).astype(np.float32), axis=1))
return batch_x, batch_y, batch_pad
def compute_score(root_dir, clf):
""" Test model with all images in given directory
"""
fvecs = []
true_classes = []
for fn in os.listdir(root_dir):
assert len(fn.split('_')) > 1, 'Invalid filename'
fname = os.path.join(root_dir, fn)
fvecs.append(get_feature_vector(fname))
true_classes.append(fn.split('_')[0])
print('Accuracy:', clf.score(fvecs, true_classes))
def classify_image(image_path, clf):
""" Classify given image
"""
fvecs = [get_feature_vector(image_path)]
print(clf.predict(fvecs), clf.predict_proba(fvecs))
