def main(img_file: str):
#tmp_path = Config.get('tmp_train_path')
tmp_path = "F:\\School\\magic-mirror\\backend\\api\\MirrorOfErised\\MirrorOfErised\\wwwroot\\images"
#tmp_path = os.path.join( "F:", "School", "magic-mirror","backend","api","MirrorOfErised","MirrorOfErised","wwwroot","images" )
#train_data_path = Config.get('train_data_path')
train_data_path = "F:\\School\\magic-mirror\\backend\\api\\MirrorOfErised\\MirrorOfErised\\wwwroot\\images\\test"
if tmp_path is None:
raise ('No temp path assigned')
full_image_path = f'{tmp_path}\\{img_file}'
#full_image_path=pathlib.Path(tmp_path, img_file)
image = Image.open(full_image_path)
image = utils.prepare_image(
image) # Check if gray scaled and resize, rotate, flip if needed
if image is None:
print('NOK')
exit(0)
detector = MTCNN()
face = detector.forward(image) # Check for faces
if face is None:
print('NOK')
exit(0)
try:
os.renames(full_image_path,
f'{train_data_path}\\{img_file.split("_")[1]}\\{img_file}')
except:
print('OK - Failed to move file')
print('OK')
def main(img_file: str):
tmp_path = Config.get('tmp_train_path')
train_data_path = Config.get('train_data_path')
if tmp_path is None:
raise ('No temp path assigned')
full_image_path = f'{tmp_path}/{img_file}'
image = Image.open(full_image_path)
image = utils.prepare_image(
image) # Check if gray scaled and resize, rotate, flip if needed
if image is None:
print('NOK')
exit(0)
detector = MTCNN()
face = detector.forward(image) # Check for faces
if face is None:
print('NOK')
exit(0)
try:
os.renames(full_image_path,
f'{train_data_path}/{img_file.split("_")[1]}/{img_file}')
except:
print('OK - Failed to move file')
print('OK')
def predict_one(image_file):
# Train a model first with train.py
facial_recognition = load_model()
detector = MTCNN()
image = Image.open(image_file)
image = utils.prepare_image(image)
face = detector.forward(image)
prediction = facial_recognition.predict(face)
print('Prediction:', prediction)
def main(test_dir):
# Train a model first with train.py
facial_recognition = load_model()
test_images = utils.load_images(test_dir)
X_test = []
y_test = []
detector = MTCNN()
for label, images in test_images.items():
for image in images:
img_file = Image.open(image)
img_file = utils.prepare_image(img_file)
face = detector.forward(img_file)
X_test.append(face)
y_test.append(label)
y_pred = facial_recognition.predict(X_test)
y_pred = np.array(y_pred)
print('Report', metrics.classification_report(y_test, y_pred))
print('Accuracy:', metrics.accuracy_score(y_test, y_pred))
