Python 3 利用 Dlib 实现摄像头实时人脸识别

# Copyright (C) 2020 coneypo

# SPDX-License-Identifier: MIT

# Author:   coneypo

# Blog:     http://www.cnblogs.com/AdaminXie

# GitHub:   

# Mail:     [e-mail]

# 从人脸图像文件中提取人脸特征存入 "features_all.csv" / Extract features from images and save into "features_all.csv"

import os

import dlib

from skimage import io

import csv

import numpy as np

# 要读取人脸图像文件的路径 / Path of cropped faces

path_images_from_camera = "data/data_faces_from_camera/"

# Dlib 正向人脸检测器 / Use frontal face detector of Dlib

detector = dlib.get_frontal_face_detector()

# Dlib 人脸 landmark 特征点检测器 / Get face landmarks

predictor = dlib.shape_predictor('data/data_dlib/shape_predictor_68_face_landmarks.dat')

# Dlib Resnet 人脸识别模型，提取 128D 的特征矢量 / Use Dlib resnet50 model to get 128D face descriptor

face_reco_model = dlib.face_recognition_model_v1("data/data_dlib/dlib_face_recognition_resnet_model_v1.dat")

# 返回单张图像的 128D 特征 / Return 128D features for single image

# Input:    path_img           

# Output:   face_descriptor    

def return_128d_features(path_img):

    img_rd = io.imread(path_img)

    faces = detector(img_rd, 1)

    print("%-40s %-20s" % ("检测到人脸的图像 / Image with faces detected:", path_img), '\n')

    # 因为有可能截下来的人脸再去检测，检测不出来人脸了, 所以要确保是 检测到人脸的人脸图像拿去算特征

    # For photos of faces saved, we need to make sure that we can detect faces from the cropped images

    if len(faces) != 0:

        shape = predictor(img_rd, faces[0])

        face_descriptor = face_reco_model.compute_face_descriptor(img_rd, shape)

    else:

        face_descriptor = 0

        print("no face")

    return face_descriptor

# 返回 personX 的 128D 特征均值 / Return the mean value of 128D face descriptor for person X

# Input:    path_faces_personX       

# Output:   features_mean_personX    

def return_features_mean_personX(path_faces_personX):

    features_list_personX = []

    photos_list = os.listdir(path_faces_personX)

    if photos_list:

        for i in range(len(photos_list)):

            # 调用 return_128d_features() 得到 128D 特征 / Get 128D features for single image of personX

            print("%-40s %-20s" % ("正在读的人脸图像 / Reading image:", path_faces_personX + "/" + photos_list[i]))

            features_128d = return_128d_features(path_faces_personX + "/" + photos_list[i])

            # 遇到没有检测出人脸的图片跳过 / Jump if no face detected from image

            if features_128d == 0:

                i += 1

            else:

                features_list_personX.append(features_128d)

    else:

        print("文件夹内图像文件为空 / Warning: No images in " + path_faces_personX + '/', '\n')

    # 计算 128D 特征的均值 / Compute the mean

    # personX 的 N 张图像 x 128D -> 1 x 128D

    if features_list_personX:

        features_mean_personX = np.array(features_list_personX).mean(axis=0)

    else:

        features_mean_personX = np.zeros(128, dtype=int, order='C')

    print(type(features_mean_personX))

    return features_mean_personX

# 获取已录入的最后一个人脸序号 / Get the order of latest person

person_list = os.listdir("data/data_faces_from_camera/")

person_num_list = []

for person in person_list:

    person_num_list.append(int(person.split('_')[-1]))

person_cnt = max(person_num_list)

with open("data/features_all.csv", "w", newline="") as csvfile:

    writer = csv.writer(csvfile)

    for person in range(person_cnt):

        # Get the mean/average features of face/personX, it will be a list with a length of 128D

        print(path_images_from_camera + "person_" + str(person + 1))

        features_mean_personX = return_features_mean_personX(path_images_from_camera + "person_" + str(person + 1))

        writer.writerow(features_mean_personX)

        print("特征均值 / The mean of features:", list(features_mean_personX))

        print('\n')

    print("所有录入人脸数据存入 / Save all the features of faces registered into: data/features_all.csv")