1.下载facenet模型
facenet须要通过大量的运算,所以间接应用他人训练过的权重
链接：https://pan.baidu.com/s/1CGyz...
提取码：pew0
2. 应用的工具
1. facenet 预训练模型,用于获取人脸特色2. mtcnn 用于人脸检测并返回对应参数3. 代码实现
1. 加载模型
from keras.models import load_modelfrom mtcnn import MTCNN# 加载mtcnndector = MTCNN()# 加载facenet模型model = load_model('facenet_keras.h5')2.依据图片检测人脸
def extract_face(detector, file_name, required_size=(160, 160)): # detector:mtcnn 对象,file_name:图片门路 # 加载图片 image = Image.open(file_name) # 将图片转换黑白 image = image.convert('RGB') # 将图片转换为 array pixels = asarray(image) # 应用 mtcnn 检测人脸 results = detector.detect_faces(pixels) # 返回人脸地位信息 x1, y1, width, height = results[0]['box'] x1, y1 = abs(x1), abs(y1) x2, y2 = x1 + width, y1 + height # 从图片剪裁人脸 face = pixels[y1:y2, x1:x2] # 将剪裁的人脸矩阵化 image = Image.fromarray(face) # 调整检测人脸图片大小 image = image.resize(required_size) # 调整矩阵形态 face_array = asarray(image).reshape(1, 160, 160, 3) return face_array3.应用facenet 提取人脸特色
def pre_process(x): if x.ndim == 4: axis = (1, 2, 3) size = x[0].size elif x.ndim == 3: axis = (0, 1, 2) size = x.size else: raise ValueError('Dimension should be 3 or 4') mean = np.mean(x, axis=axis, keepdims=True) std = np.std(x, axis=axis, keepdims=True) std_adj = np.maximum(std, 1.0 / np.sqrt(size)) y = (x - mean) / std_adj return ydef get_embedding(model, img): face_img = pre_process(img) pre = model.predict(face_img) pre = l2_normalize(np.concatenate(pre)) pre = np.reshape(pre, [128]) return pre4.人脸特色比照
当从输出的两张图片中获取到人脸特色后,须要将两张图片的特色进行比照,具体能够参考吴恩达老师的深度学习视频
def face_distance(face_encodings, face_to_compare): if len(face_encodings) == 0: return np.empty((0)) return np.linalg.norm(face_encodings - face_to_compare, axis=1)4.残缺代码
import osfrom keras.models import load_modelfrom numpy import expand_dimsimport numpy as npfrom face_detector import extract_facefrom datetime import datetimefrom mtcnn import MTCNNimport utilsfrom PIL import Imagefrom numpy import asarraydef pre_process(x): if x.ndim == 4: axis = (1, 2, 3) size = x[0].size elif x.ndim == 3: axis = (0, 1, 2) size = x.size else: raise ValueError('Dimension 3 or 4') mean = np.mean(x, axis=axis, keepdims=True) std = np.std(x, axis=axis, keepdims=True) std_adj = np.maximum(std, 1.0 / np.sqrt(size)) y = (x - mean) / std_adj return ydef l2_normalize(x, axis=-1, epsilon=1e-10): output = x / np.sqrt(np.maximum(np.sum(np.square(x), axis=axis, keepdims=True), epsilon)) return outputdef get_embedding(model, img): face_img = pre_process(img) pre = model.predict(face_img) pre = l2_normalize(np.concatenate(pre)) pre = np.reshape(pre, [128]) return predef extract_face(detector, file_name, required_size=(160, 160)): # 加载图片 image = Image.open(file_name) # 将图片转换黑白 image = image.convert('RGB') # 将图片转换为 array pixels = asarray(image) # 应用 mtcnn 检测人脸 results = detector.detect_faces(pixels) # 返回人脸地位信息 x1, y1, width, height = results[0]['box'] x1, y1 = abs(x1), abs(y1) x2, y2 = x1 + width, y1 + height # 从图片剪裁人脸 face = pixels[y1:y2, x1:x2] # 将剪裁的人脸矩阵化 image = Image.fromarray(face) # 调整检测人脸图片大小 image = image.resize(required_size) # 调整矩阵形态 face_array = asarray(image).reshape(1, 160, 160, 3) return face_arraydef face_distance(face_encodings, face_to_compare): if len(face_encodings) == 0: return np.empty((0)) return np.linalg.norm(face_encodings - face_to_compare, axis=1)def verify_demo(): model = load_model('facenet_keras.h5') dector = MTCNN() org_list = [] # 图片门路,批改成本人的图片门路 image1_path = 'images\camera_0.jpg' pixels = extract_face(dector, image1_path) embedding1 = get_embedding(model, pixels) org_list.append(embedding1) # 图片门路,批改成本人的图片门路 image2_path = 'vertify\younes.jpg' pixels1 = extract_face(dector, image2_path) embedding2 = get_embedding(model, pixels1) dist = face_distance(org_list, embedding2) print(dist) if dist < 0.6: print("欢送 " + str('ounes') + "回家！") else: print("教训证，您与" + str('ounes') + "不符！")if __name__ == '__main__': verify_demo() # check_who_is()参考:
吴恩达课后编程作业】Course 4 -卷积神经网络 - 第周围作业
如何在 Keras 中应用 FaceNet 开发人脸识别零碎