dlibdotnet 人脸相似度源代码_OpenCV3与深度学习实例：Dlib+VGG Face实现两张脸部图像相似度比较...

实现思路：

1、使用Dlib识别并提取脸部图像

2、使用VGG Face模型提取脸部特征

3、使用余弦相似度算法比较两张脸部图像的特征

代码如下：

import time

import numpy as np

import sklearn

import sklearn.metrics.pairwise as pw

import cv2

import dlib

prototxt = 'datas/models/caffe/vgg-face/vgg_face_caffe/vgg_face_caffe/VGG_FACE_deploy.prototxt'

caffemodel = 'datas/models/caffe/vgg-face/vgg_face_caffe/vgg_face_caffe/VGG_FACE.caffemodel'

dlib_model = 'datas/models/dlib/shape_predictor_68_face_landmarks.dat'

detector = dlib.get_frontal_face_detector()

predictor = dlib.shape_predictor(dlib_model)

net = cv2.dnn.readNetFromCaffe(prototxt, caffemodel)

faces1 = get_faces('datas/images/face-tests/fanbb/f16.jpg')

faces2 = get_faces('datas/images/faces/fanbb.jpg')

for i,face in