黑猴子的家:Python的开源人脸识别,face_recogni
2019-01-23 本文已影响83人
黑猴子的家
1、GitHub人脸识别库
网址
https://github.com/ageitgey/face_recognition#face-recognition
2、简介
该库可以通过python或者命令行即可实现人脸识别的功能。使用dlib深度学习人脸识别技术构建,在户外脸部检测数据库基准(Labeled Faces in the Wild)上的准确率为99.38%。
在github上有相关的链接和API文档。
3、运行 Anaconda Prompt
4、配置国内镜像
https://mirrors.tuna.tsinghua.edu.cn/help/anaconda/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/
conda config --set show_channel_urls yes
5、创建python环境和第三方库
(base)C:\Users\Administrator>conda create –n py36 python=3.6
(base)C:\Users\Administrator>y
(base)C:\Users\Administrator>activate py36
(py36)C:\Users\Administrator>pip install dlib (直接安装会报错)
(py36)C:\Users\Administrator>pip install dlib-19.7.0-cp36-cp36m-win_amd64.whl
(py36)C:\Users\Administrator>pip install face_recognition-1.2.3-py2.py3-none-any.whl
(py36)C:\Users\Administrator>pip install opencv_python
(py36)C:\Users\Administrator>conda install spyder
(py36)C:\Users\Administrator>conda install gevent
(py36)C:\Users\Administrator>pip install freetype-py
(py36)C:\Users\Administrator>conda list
(py36)C:\Users\Administrator>spyder
去https://pypi.org/project/dlib/#history 直接下一个支持python3.6 且版本号大于19.4的dlib,格式为whl 同时也下载了一个face-recognition.whl
6、识别人脸
code
# -*- coding: utf-8 -*-
import face_recognition
import cv2
# 读取图片
#img = face_recognition.load_image_file("/Users/z/Desktop/group_face2/teacherbanner.jpg")
img = face_recognition.load_image_file("./22.png")
# 得到人脸坐标
face_locations = face_recognition.face_locations(img)
print(face_locations)
# 显示原始图片
img = cv2.imread("./22.png")
cv2.namedWindow("original")
cv2.imshow("original", img)
# 遍历每个人脸
faceNum = len(face_locations)
for i in range(0, faceNum):
top = face_locations[i][0]
right = face_locations[i][1]
bottom = face_locations[i][2]
left = face_locations[i][3]
start = (left, top)
end = (right, bottom)
color = (247, 230, 16)
thickness = 2
cv2.rectangle(img, start, end, color, thickness)
# 显示识别后的图片
cv2.namedWindow("recognition")
cv2.imshow("recognition", img)
cv2.waitKey(0)
cv2.destroyAllWindows()
7、人脸识别
# -*- coding: utf-8 -*-
import face_recognition
import cv2
from gevent import os
import freetype
import copy
class ChineseTextUtil(object):
def __init__(self, ttf):
self._face = freetype.Face(ttf)
def draw_text(self, image, pos, text, text_size, text_color):
'''
使用ttf字体库中的字体设置姓名
:param image: 用于将text生成在某个image图像上
:param pos: 画text的位置
:param text: unicode编码的text
:param text_size: 字体大小
:param text_color:字体颜色
:return: 返回位图
'''
self._face.set_char_size(text_size * 64)
metrics = self._face.size
ascender = metrics.ascender / 64.0
# descender = metrics.descender / 64.0
# height = metrics.height / 64.0
# linegap = height - ascender + descender
ypos = int(ascender)
#unicode = ('utf-8','unicode')
#if not isinstance(text, unicode):
#text = text.decode('utf-8')
img = self.string_2_bitmap(image, pos[0], pos[1], text, text_color)
return img
def string_2_bitmap(self, img, x_pos, y_pos, text, color):
'''
将字符串绘制为图片
:param x_pos: text绘制的x起始坐标
:param y_pos: text绘制的y起始坐标
:param text: text的unicode编码
:param color: text的RGB颜色编码
:return: 返回image位图
'''
prev_char = 0
pen = freetype.Vector()
pen.x = x_pos << 6 # div 64
pen.y = y_pos << 6
hscale = 1.0
matrix = freetype.Matrix(int(hscale) * 0x10000, int(0.2 * 0x10000), int(0.0 * 0x10000), int(1.1 * 0x10000))
cur_pen = freetype.Vector()
pen_translate = freetype.Vector()
image = copy.deepcopy(img)
for cur_char in text:
self._face.set_transform(matrix, pen_translate)
self._face.load_char(cur_char)
kerning = self._face.get_kerning(prev_char, cur_char)
pen.x += kerning.x
slot = self._face.glyph
bitmap = slot.bitmap
cur_pen.x = pen.x
cur_pen.y = pen.y - slot.bitmap_top * 64
self.draw_ft_bitmap(image, bitmap, cur_pen, color)
pen.x += slot.advance.x
prev_char = cur_char
return image
def draw_ft_bitmap(self, img, bitmap, pen, color):
'''
draw each char
:param bitmap: 位图
:param pen: 画笔
:param color: 画笔颜色
:return: 返回加工后的位图
'''
x_pos = pen.x >> 6
y_pos = pen.y >> 6
cols = bitmap.width
rows = bitmap.rows
glyph_pixels = bitmap.buffer
for row in range(rows):
for col in range(cols):
if glyph_pixels[row * cols + col] != 0:
img[y_pos + row][x_pos + col][0] = color[0]
img[y_pos + row][x_pos + col][1] = color[1]
img[y_pos + row][x_pos + col][2] = color[2]
if __name__ == '__main__':
# 读取图片识别样例
face_file_list = []
names_list = []
face_encoding_list = []
rootdir = './'
list = os.listdir(rootdir)
for i in range(0, len(list)):
path = os.path.join(rootdir, list[i])
if os.path.isfile(path) and ".jpg" in list[i]:
face_file_list.append(rootdir + list[i])
print(list[i][:-4])
names_list.append(list[i][:-4])
for path in face_file_list:
print(path)
face_image = face_recognition.load_image_file(path)
face_encoding = face_recognition.face_encodings(face_image)[0]
face_encoding_list.append(face_encoding)
# 初始化一些变量用于,面部位置,编码,姓名等
face_locations = []
face_encodings = []
face_names = []
process_this_frame = True
video_capture = cv2.VideoCapture(0)
while True:
# 得到当前摄像头拍摄的每一帧
ret, frame = video_capture.read()
# 缩放当前帧到4分支1大小,以加快识别进程的效率
small_frame = cv2.resize(frame, (0, 0), fx=0.25, fy=0.25)
# 每次只处理当前帧的视频,以节省时间
if process_this_frame:
# 在当前帧中,找到所有的面部的位置以及面部的编码
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:
# 找到能够与已知面部匹配的面部
match = face_recognition.compare_faces(face_encoding_list, face_encoding, 0.6)
name = "Unknown"
for i in range(0, len(match)):
if match[i]:
name = names_list[i]
face_names.append(name)
process_this_frame = not process_this_frame
# 显示结果
for (top, right, bottom, left), name in zip(face_locations, face_names):
# 将刚才缩放至4分支1的帧恢复到原来大小,并得到与每一个面部与姓名的映射关系
top *= 4
right *= 4
bottom *= 4
left *= 4
# 在脸上画一个框框
cv2.rectangle(frame, (left, top), (right, bottom), (0, 0, 255), 2)
# 在框框的下边画一个label用于显示姓名
#cv2.rectangle(frame, (left, bottom - 35), (right, bottom), (0, 0, 255), cv2.cv.CV_FILLED)
cv2.rectangle(frame, (left, bottom - 35), (right, bottom), (0, 0, 255), 1)
font = cv2.FONT_HERSHEY_DUPLEX
# 在当前帧中显示我们识别的结果
color_ = (255, 255, 255)
pos = (left + 6, bottom - 6)
text_size = 24
# 使用自定义字体
ft = ChineseTextUtil('wqy-zenhei.ttc')
image = ft.draw_text(frame, pos, name, text_size, color_)
cv2.imshow('VideoZH', image)
# cv2.putText(frame, name, (left + 6, bottom - 6), font, 1.0, (255, 255, 255), 1)
# cv2.imshow('Video', frame)
# 按q退出
if cv2.waitKey(1) & 0xFF == ord('q'):
break
# 释放资源
video_capture.release()
cv2.destroyAllWindows()