"""
Handles all requests relevant to the verification service of the API.
"""
from image_processing.sample_extract import TextExtractor
from image_preprocessing.face_manager import FaceDetector
from verification.text_verify import TextVerify
from verification.face_verify import FaceVerify
from flask import jsonify, request, Blueprint
from hutts_utils.hutts_logger import logger
from hutts_utils.image_handling import grab_image
from hutts_utils.pypath import correct_path
from pathlib import Path
import os
__authors__ = "Nicolai van Niekerk, Stephan Nell, Marno Hermann"
__copyright__ = "Copyright 2017, Java the Hutts"
__license__ = "BSD"
__maintainer__ = "Nicolai van Niekerk"
__email__ = "nicvaniek@gmail.com"
__status__ = "Development"
verify = Blueprint('verify', __name__)
# Constants path to trained data for Shape Predictor.
CURRENT_LOCATION = os.path.abspath(os.path.dirname(__file__))
SHAPE_PREDICTOR_PATH = correct_path(Path(
Path(CURRENT_LOCATION),
Path(CURRENT_LOCATION).parent,
'image_preprocessing',
'trained_data',
'shape_predictor_face_landmarks.dat'
)
)
FACE_RECOGNITION_PATH = correct_path(Path(
Path(CURRENT_LOCATION),
Path(CURRENT_LOCATION).parent,
'image_preprocessing',
'trained_data',
'dlib_face_recognition_resnet_model_v1.dat'
))
[docs]@verify.route('/verifyID', methods=['POST'])
def verify_id():
"""
Sample function to return a match percentage of an ID image and
provided personal information and picture of face.
URL: http://localhost:5000/verifyID
"""
image_of_id, face = receive_faces(match_face=True)
entered_details = receive_details()
is_match, distance = match_faces(image_of_id, face)
extracted_text, preferences = manage_text_extractor(image_of_id)
text_match_percentage, text_match, is_pass = manage_text_verification(preferences, extracted_text, entered_details)
logger.info("Preparing Results...")
result = {
# text verification contributes to 40% of the total and face likeness for 60%
"total_match": text_match_percentage*0.4 + distance*0.6,
"text_match": text_match,
"face_match": distance,
"is_match": is_match,
"is_pass": is_pass
}
return jsonify(result)
[docs]@verify.route('/verifyFaces', methods=['POST'])
def verify_faces():
"""
Sample function to return a match percentage of an ID face image and
picture of face
URL: http://localhost:5000/verifyFaces
"""
image_of_id, face = receive_faces(match_face=True)
(is_match, distance) = match_faces(image_of_id, face)
logger.info("Preparing Results...")
result = {
"face_match": distance,
"is_match": is_match,
}
return jsonify(result)
[docs]@verify.route('/verifyInfo', methods=['POST'])
def verify_info():
"""
Sample function to return a match percentage of an ID image and
provided personal information
"""
image_of_id, _ = receive_faces(match_face=False)
entered_details = receive_details()
extracted_text, preferences = manage_text_extractor(image_of_id)
text_match_percentage, text_match, is_pass = manage_text_verification(preferences, extracted_text, entered_details)
logger.info("Preparing Results...")
result = {
"text_match": text_match,
"is_pass": is_pass,
"text_match_percentage": text_match_percentage
}
return jsonify(result)
[docs]def match_faces(image_of_id, face):
"""
This function receives two images that receive images of faces that need to be verified.
It is expected that an image of an ID and an image of a Profile picture will be received.
Even if the expected images are not received the function will still apply a best effort solution.
Args:
image_of_id (:obj:'OpenCV image'): An image of an ID that contains a face that needs to be verified
face (:obj:'OpenCV image'): A image of a face that needs to be verified
Returns:
bool: Represent if two faces indeed match True if distance calculated is
below a threshold value. False if the distance calculated is above threshold value.
float: Return Euclidean distance between the vector representation
of the two faces.
"""
# Extract face
face_detector = FaceDetector(SHAPE_PREDICTOR_PATH)
extracted_face1 = face_detector.extract_face(image_of_id)
extracted_face2 = face_detector.extract_face(face)
# Verify faces
face_verifier = FaceVerify(SHAPE_PREDICTOR_PATH, FACE_RECOGNITION_PATH)
return face_verifier.verify(extracted_face1, extracted_face2)
[docs]def receive_faces(match_face=True):
"""
This function receives faces/ID from request flask handler.
The function checks for multiple means of receiving the faces/ID. These include
Receiving image by file path
Receiving image by URL
Receiving image by file Stream
It is expected that an image of a face and an image of an ID will be sent.
However, if the order is not followed that system will still be able to return the best effort result without
loss of accuracy.
Args:
match_face (bool): Indicates if additional profile of a face should be extracted.
If an additional face should not be extracted simply return the ID image.
Returns:
image_of_id (:obj:'OpenCV image'): An image of a ID.
face (:obj:'OpenCV image'): An image of a face. If match_face is set to True
"""
data = {"success": False}
# Get id image as numpy array
# Check to see if an image was uploaded.
if request.files.get("id_img", None) is not None:
# Grab the uploaded image.
image_of_id = grab_image(stream=request.files["id_img"])
# Otherwise, assume that a URL was passed in.
else:
# Grab the URL from the request.
url = request.args.get("url", None)
# If the URL is None, then return an error.
if url is None:
data["error"] = "No URL provided."
return jsonify(data)
# Load the image and convert.
image_of_id = grab_image(url=url)
if not match_face:
return image_of_id
# Get face image as numpy array
# Check to see if an image was uploaded.
if request.files.get("face_img", None) is not None:
# Grab the uploaded image.
face = grab_image(stream=request.files["face_img"])
# Otherwise, assume that a URL was passed in.
else:
# Grab the URL from the request.
url = request.args.get("url", None)
# If the URL is None, then return an error.
if url is None:
data["error"] = "No URL provided."
return jsonify(data)
# Load the image and convert.
face = grab_image(url=url)
return image_of_id, face
[docs]def manage_text_verification(preferences, extracted_text, entered_details):
"""
The function manages the preparation before text verification and result of the text verification it self
Management includes preparing logger functionality and controlling match percentages and messages.
Args:
preferences (list): A list of preferences. Containing details about logger functionality.
extracted_text (JSON object): Contains text extracted from ID image.
entered_details (dict): Dictionary containing information that needs to be verified
Returns:
text_match_percentage (float): Value representing the accuracy with which the entered details
matches that of the extracted text.
text_match (dict): Contains intermediate match percentages for different details.
For Example the match percentage between just the date_of_birth fields
is_pass (bool): Indicates if text_verification passes.
If match percentage is above threshold provided in preferences then verification passes.
If match percentage is below threshold value provided in preferences then verification fails.
"""
text_verifier = TextVerify()
verbose_verify = preferences['verbose_verify']
logger.debug('%s text verifiction requested' % ('Verbose' if verbose_verify else 'Non-verbose'))
(is_pass, text_match) = text_verifier.verify(extracted_text, entered_details, verbose=verbose_verify)
# Check if we are working with verbose output for text verification
text_match_percentage = text_match if not verbose_verify else text_match['total']
return text_match_percentage, text_match, is_pass
[docs]def receive_details():
"""
This function receives the details that need to be verified from the flask handler.
Returns:
entered_details (dict): Details that need to be verified with that extracted from image.
"""
entered_details = {
"names": request.form['names'],
"surname": request.form['surname'],
"identity_number": request.form['idNumber'],
"nationality": request.form['nationality'],
"country_of_birth": request.form['cob'],
"status": request.form['status'],
"sex": request.form['gender'],
"date_of_birth": request.form['dob']
}
return entered_details