Python Build face recognition
Published: 2019-05-27

Face recognition is very simple for human beings. If it is for machines, how can a face recognition be constructed?

The main steps are as follows:

(1) Define label code.In the input training data, labels are expressed in words, but we need numbers to train the system.

(2) Extracting ROI attribute values and label encoders from each map.

(3) Load the face cascade file.

(4) generating a local binary pattern histogram face recognition.

(5) Training face recognition with training set data.

(6) Load the path where the test data is located, read the batch data into the memory, and use the human face cascade file to determine the position of the human face.

(7) For each person's face ROI, run the face recognition to convert the label into a word.

(8) Typing classified text directly on the picture and displaying.

(9) Close cv2.

Source code is as follows:

import osimport cv2import numpy as npfrom sklearn import preprocessing# Class to handle tasks related to label encodingclass LabelEncoder(object):    # Method to encode labels from words to numbers    def encode_labels(self, label_words):        self.le = preprocessing.LabelEncoder()    # Convert input label from word to number    def word_to_num(self, label_word):        print("[label_word]",[label_word])        print("self.le.transform([label_word])[0]",self.le.transform([label_word])[0])        print("int(self.le.transform([label_word])[0])",int(self.le.transform([label_word])[0]))        return int(self.le.transform([label_word])[0])    # Convert input label from number to word    def num_to_word(self, label_num):        return self.le.inverse_transform([label_num])[0]# Extract images and labels from input pathdef get_images_and_labels(faceCascade,input_path):    label_words = []    # Iterate through the input path and append files    for root, dirs, files in os.walk(input_path):        for filename in (x for x in files if x.endswith('.jpg')):            filepath = os.path.join(root, filename)            label_words.append(filepath.split('\\')[-2])    print("label_words",label_words)    print("filepath",filepath)    # Initialize variables    images = []    le = LabelEncoder()    le.encode_labels(label_words)    print("le",le)    labels = []    # Parse the input directory    for root, dirs, files in os.walk(input_path):        for filename in (x for x in files if x.endswith('.jpg')):            filepath = os.path.join(root, filename)            # Read the image in grayscale format            image = cv2.imread(filepath, 0)            # Extract the label            name = filepath.split('\\')[-2]            # Perform face detection            faces = faceCascade.detectMultiScale(image, 1.1, 2, minSize=(100,100))            # Iterate through face rectangles            # print(faces)            for (x, y, w, h) in faces:                images.append(image[y:y+h, x:x+w])                labels.append(le.word_to_num(name))                print("x,y,w,h",x,y,w,h,"filepath=",filepath)            print("labels=",labels)    return images, labels, leclass my_face_reconginizer:    def __init__(self,cascade_path = r"cascade_files\haarcascade_frontalface_alt.xml",                      path_train = r'faces_dataset\train',path_test = r'faces_dataset\Test'): self.facecade = cv2.cascadeclassifier (cascade _ path) # face cascade file.        self.path_train=path_train        self.path_test=path_test    def Recongizer_train(self):        self.images, self.labels, self.le = get_images_and_labels(self.faceCascade,self.path_train)        self.recognizer = cv2.face.LBPHFaceRecognizer_create()        print( "\nTraining...")        self.recognizer.train(self.images, np.array(self.labels))    def Recongizer_Predict(self):        # Test the recognizer on unknown images        print('\nPerforming prediction on test images...')        stop_flag = False        for root, dirs, files in os.walk(self.path_test):            for filename in (x for x in files if x.endswith('.jpg')):                filepath = os.path.join(root, filename)                # Read the image                predict_image = cv2.imread(filepath, 0)                print("predict_image",predict_image)                print("filepath",filepath)                print("type(predict_image)",type(predict_image))                print("shape(predict_image)",np.shape(predict_image))                # Detect faces                faces = self.faceCascade.detectMultiScale(predict_image, 1.1,                        2, minSize=(100,100))                print("faces",faces)                # Iterate through face rectangles                for (x, y, w, h) in faces:                    # Predict the output                    predicted_index, conf = self.recognizer.predict(                            predict_image[y:y+h, x:x+w])                    # Convert to word label                    predicted_person = self.le.num_to_word(predicted_index)                    # Overlay text on the output image and display it                    cv2.putText(predict_image, 'Prediction: ' + predicted_person,                            (10,60), cv2.FONT_HERSHEY_SIMPLEX, 1, (255,255,255),3)                    cv2.imshow("Recognizing face", predict_image)                c = cv2.waitKey(0)                if c == 27:                    stop_flag = True                    break            return 0            if stop_flag:                breaktest1=my_face_reconginizer()test1.Recongizer_train()test1.Recongizer_Predict()