utils.py

import re
import os
import time
import dlib
import cv2
import yaml
import numpy as np
import pytesseract as tess

from scipy.spatial import distance


def crop_face_from_id(cv_image, weight_path="weights"):
    detector = dlib.get_frontal_face_detector()
    predictor = dlib.shape_predictor(os.path.join(weight_path, 'shape_predictor_68_face_landmarks.dat'))

    img_gray = cv2.cvtColor(cv_image, cv2.COLOR_BGR2GRAY)

    face_detector = detector(img_gray, 1)

    if len(face_detector) == 1:
        face = face_detector[0]

        landmarks = predictor(cv_image, face)

        landmark_list = []
        for p in landmarks.parts():
            landmark_list.append([p.x, p.y])
        h = landmark_list[8][1] - landmark_list[27][1]
        w = landmark_list[16][0] - landmark_list[0][0]
        (x, y) = landmark_list[33]

        cropped = cv_image.copy()
        cropped = cropped[y - int(h * 1.25): y + h//2, x - int(w * 0.75): x + int(w * 0.75)]

        return cropped

    else:
        raise RuntimeError("[Error] Cannot find ID card")


def get_idnum(cv_image):
    """extract birth year from id card by Tesseract"""
    tess.pytesseract.tesseract_cmd = r"C:\Program Files\Tesseract-OCR\tesseract.exe"
    config = ("-l kor+eng --oem 3 --psm 4 -c preserve_interword_spaces=1")
    text = tess.image_to_string(cv_image, config=config)
    try:
        text = re.compile(r"\d{6}").search(text).group()
        # text = text.replace('-', '')
            
        return text
    
    except RuntimeError:
        print("[Error] Cannot find ID card")


def is_verified_idnum(image):
    try:
        id_num = get_idnum(image)

        weights = np.array([2, 3, 4, 5, 6, 7, 8, 9, 2, 3, 4, 5])
        id_num_array = np.array(list(id_num[:-1]), np.int32)

        if int(id_num[-1]) == (11 - np.dot(weights, id_num_array.transpose()) % 11) % 11:
            return True

        else:
            return False

    except RuntimeError:
        print("[Error] Cannot find ID card")


def is_verified_age(image):
    try:
        id_num = get_idnum(image)
        year_curr = int(time.strftime('%Y', time.localtime(time.time())))
        # if id_num[6] == '1' or id_num[6] == '2':
        if not id_num[0] == '0':
            year_birth = int(id_num[:2]) + 1900
        else:
            year_birth = int(id_num[:2]) + 2000

        age = year_curr - year_birth
        if age >= 19:
            return True
        else:
            return False

    except RuntimeError:
        print("[Error] Cannot find ID card")


def eye_aspect_ratio(eye):
    a = distance.euclidean(eye[1], eye[5])
    b = distance.euclidean(eye[2], eye[4])
    c = distance.euclidean(eye[0], eye[3])
    ear = (a + b) / (2.0 * c)
    return ear


def mouth_aspect_ratio(mouth):
    a = distance.euclidean(mouth[14], mouth[18])    # 51, 59
    b = distance.euclidean(mouth[4], mouth[8])      # 53, 57
    c = distance.euclidean(mouth[12], mouth[16])    # 49, 55
    mar = (a + b) / (2.0 * c)
    return mar


def l2_norm(x, axis=1):
    norm = np.linalg.norm(x, axis=axis, keepdims=True)
    return x / norm


def get_embeddings(recognition_model, img):
    img = cv2.resize(img, (112, 112))
    img = img / 255.
    if len(img.shape) == 3:
        img = np.expand_dims(img, axis=0)
    embeds = l2_norm(recognition_model(img))
    return embeds


def load_yaml(load_path):
    """load yaml file"""
    with open(load_path, 'r') as f:
        loaded = yaml.load(f, Loader=yaml.Loader)

    return loaded


def get_ckpt_inf(ckpt_path, steps_per_epoch):
    """get ckpt information"""
    split_list = ckpt_path.split('e_')[-1].split('_b_')
    epochs = int(split_list[0])
    batchs = int(split_list[-1].split('.ckpt')[0])
    steps = (epochs - 1) * steps_per_epoch + batchs

    return epochs, steps + 1