admin 发布时间:2019-12-11 分类:分享 阅读:7105次 4 条评论
使用到了开源库 face_recognition,github地址:https://github.com/ageitgey/face_recognition
安装过程以Ubuntu及Python2.7为例:
apt-get update
apt-get install build-essential cmake apt-get install libgtk-3-dev apt-get install libboost-all-dev
#pip install dlibapt-get install python-pip apt-get install python-pip
#pip install face_recognition pip install --index https://mirrors.ustc.edu.cn/pypi/web/simple/ face_recognition
以上是插件的安装命令,安装完成后提示 Successfully,以下附上Python示例:
# coding:utf-8 import face_recognition #输入已知图片1.jpg known_image = face_recognition.load_image_file("1.jpg") #输入待识别的图片2.jpg unknown_image = face_recognition.load_image_file("2.jpg") biden_encoding = face_recognition.face_encodings(known_image)[0] unknown_encoding = face_recognition.face_encodings(unknown_image)[0] results = face_recognition.compare_faces([biden_encoding], unknown_encoding) #输出的results是一串Boolean值 print results
保存为facedemo.py文件,并在当前目录放置两张包含人脸信息的图片做对比,python facedemo.py 执行该demo。返回true表示两张图为同一人,否则匹配不通过。其中有几个API函数:
1、face_encodings 人脸解码:输入一张图片后,生成一个128维的特征向量,这是 人脸识别的依据。
2、compare_faces 人脸比对:人脸识别的核心,设置一个阈值,若两张人脸的特征向量的距离,在阈值范围之内,则认为其是同一个人,最后返回一个Boolean值的list。另外一个类似的函数是 face_distance 人脸特征向量距离,输出由Boolean对错改为数字number。
其他的函数请参考官方的API地址:https://face-recognition.readthedocs.io/en/latest/face_recognition.html
python3的face_recognition教程
安装完python3之后,继续安装pip3:
apt install python3-pip apt install cmake
再安装其他的开发插件等
pip3 install Flask # 安装 OpenCV开发包 pip3 install opencv-python # 安装最新的OpenCV 扩展 pip3 install opencv-contrib-python pip3 install matplotlib sudo apt-get update #安装依赖库 sudo apt-get install libhdf5-dev sudo apt-get install libatlas-base-dev sudo apt-get install libjasper-dev sudo apt-get install libqt4-test sudo apt-get install libqtgui4 sudo apt-get update python3 import cv2 # 检查导入成功 cv2.__version__ # 检查cv2的版本
import cv2 可能会出现报错:
Traceback (most recent call last): File "<stdin>", line 1, in <module> ImportError: /home/pi/.local/lib/python3.7/site-packages/cv2/cv2.cpython-37m-arm-linux-gnueabihf.so: undefined symbol: __atomic_fetch_add_8
需要运行:vim.tiny .bashrc
将该内容追加到文件底部:
export LD_PRELOAD=/usr/lib/arm-linux-gnueabihf/libatomic.so.1
使之生效:
source .bashrc
以下是完整简单的Python的Web实例,提供人脸注册及识别两个接口:
import os from flask import Flask, request, redirect, url_for,jsonify #from werkzeug import secure_filename from pathlib import Path import face_recognition #头像文件保存路径 UPLOAD_FOLDER = '/root/face/' #linux ALLOWED_EXTENSIONS = set(['jpg']) app = Flask(__name__) app.config['UPLOAD_FOLDER'] = UPLOAD_FOLDER @app.route('/face/register', methods=['POST']) def allowed_file(filename): return '.' in filename and filename.rsplit('.', 1)[1] in ALLOWED_EXTENSIONS def get_file_extension(path): return os.path.splitext(path)[1] def get_file_name(path): return os.path.splitext(path)[0] # 人脸照片上传 @app.route('/face/uploader', methods=['POST']) def face_uploader(): userName = request.values.get('userName') file = request.files['file'] if(not userName or not file): return jsonify({'code': 100,'msg':'error'}) filepath = app.config['UPLOAD_FOLDER'] if file and allowed_file(file.filename): #filename = secure_filename(file.filename) # 中文异常 dirpath = app.config['UPLOAD_FOLDER'] if not Path(dirpath).is_dir(): os.mkdir(dirpath) filepath = os.path.join(dirpath, userName + get_file_extension(file.filename)) file.save(filepath) return jsonify({'code': 200,'msg':'success','data':{'userName':userName,'filePath':filepath}}) # 上传照片识别 与上接口的人脸对比并返回名称 @app.route('/face/recog', methods=['POST']) def face_recog(): file = request.files['file'] if(file is None): return jsonify({'code': 100,'msg':'error'}) filepath = '' if file and allowed_file(file.filename): #filename = secure_filename(file.filename) dirpath = app.config['UPLOAD_FOLDER'] + "temp/" if not Path(dirpath).is_dir(): os.mkdir(dirpath) filepath = os.path.join(dirpath, file.filename) file.save(filepath) print("temp filepath: " + filepath) unknown_image_file = face_recognition.load_image_file(filepath) unknown_image_encodeds = face_recognition.face_encodings(unknown_image_file) if len(unknown_image_encodeds) == 0: print("the upload img no face") return jsonify({'code': 101,'msg':'no face'}) face_person_name = '' ## 查找已有的头像文件 #names = [name for name in os.listdir(app.config['UPLOAD_FOLDER']) # if os.path.isfile(os.path.join(app.config['UPLOAD_FOLDER'], name)) and # name.endswith('.jpg')] images = os.listdir(app.config['UPLOAD_FOLDER']) # 遍历每张图像 for image in images: current_path = os.path.join(app.config['UPLOAD_FOLDER'], image) if not image.endswith('.jpg') or not os.path.isfile(current_path): continue print("current_image filepath: " + current_path) current_image = face_recognition.load_image_file(current_path) # 将加载图像编码为特征向量 current_image_encodeds = face_recognition.face_encodings(current_image) if len(current_image_encodeds) == 0: print("no face") continue # 将你的图像和图像对比,看是否为同一人 result = face_recognition.compare_faces([unknown_image_encodeds[0]],current_image_encodeds[0]) if result[0] == True: print("Matched: " + image) face_person_name = get_file_name(image) else: print("Not matched: " + image) ## 查找已有的头像文件 #names = [name for name in os.listdir(app.config['UPLOAD_FOLDER']) # if os.path.isfile(os.path.join(app.config['UPLOAD_FOLDER'], name)) and # name.endswith('.jpg')] #known_face_list = [] #for filename in names: # face_image = # face_recognition.load_image_file(os.path.join(app.config['UPLOAD_FOLDER'], # filename)) # face_encoding = face_recognition.face_encodings(face_image)[0] # known_face_list.append(face_encoding) #face_return = #face_recognition.compare_faces(known_face_list,unknown_image_file) return jsonify({'code': 200,'msg':'success','data':face_person_name}) if __name__ == "__main__": # 将host设置为0.0.0.0,端口5000 则外网用户也可以访问到这个服务 app.run(host="0.0.0.0", port=5000, debug=True)
启动后通过5000端口访问:
接口1进行人脸照片上传,地址:http://*:5000/face/uploader ,使用form-data上传图片及人名。
接口2人脸对比并返回人名,地址:http://*:5000/face/recog ,使用form-data上传待识别的图片。
结合摄像头使用实时的人脸识别
下面的demo是python摄像头实时画面:
import cv2 cap = cv2.VideoCapture(0) while True: ret,frame = cap.read() cv2.imshow('Video',frame) c = cv2.waitKey(1) if c == 27: break cap.release() cv2.destroyAllWindows()
下面的demo是python根据摄像头实时画面识别人脸:
# -*- coding: utf-8 -*- import face_recognition import cv2 video_capture = cv2.VideoCapture(0) person_img = face_recognition.load_image_file("personname.jpg") person_face_encoding = face_recognition.face_encodings(person_img)[0] face_locations = [] face_encodings = [] face_names = [] process_this_frame = True while True: ret, frame = video_capture.read() small_frame = cv2.resize(frame, (0, 0), fx=0.25, fy=0.25) if process_this_frame: face_locations = face_recognition.face_locations(small_frame) face_encodings = face_recognition.face_encodings(small_frame, face_locations) face_names = [] for face_encoding in face_encodings: match = face_recognition.compare_faces([person_face_encoding], face_encoding) if match[0]: name = "personname" else: name = "unknown" face_names.append(name) process_this_frame = not process_this_frame for (top, right, bottom, left), name in zip(face_locations, face_names): top *= 4 right *= 4 bottom *= 4 left *= 4 cv2.rectangle(frame, (left, top), (right, bottom), (0, 0, 255), 2) cv2.rectangle(frame, (left, bottom - 35), (right, bottom), (0, 0, 255), 2) font = cv2.FONT_HERSHEY_DUPLEX cv2.putText(frame, name, (left+6, bottom-6), font, 1.0, (255, 255, 255), 1) cv2.imshow('Video', frame) if cv2.waitKey(1) & 0xFF == ord('q'): break video_capture.release() cv2.destroyAllWindows()
上一篇:曾经你像风
发布于 2020-05-14 21:00:40 回复该评论
发布于 2020-04-20 21:23:45 回复该评论
发布于 2020-03-02 00:30:24 回复该评论
发布于 2020-02-27 10:05:55 回复该评论
发表评论:
◎欢迎您的参与讨论。