python简单的人脸识别系统（PCA+逻辑回归）

python简单的人脸识别系统（PCA+逻辑回归）

数据集：ORL人脸库
特征提取：PCA降维，将112*92降成30
分类器：逻辑回归
**

代码：

from tkinter import *
from tkinter.tix import Tk,Control,ComboBox#升级的控件组包
from tkinter.messagebox import showinfo,showwarning,showerror#各种类型的提示框
from tkinter import filedialog
from PIL import Image,ImageTk
from tkinter.messagebox import *

import os
import operator
from numpy import *
import cv2
from sklearn.decomposition import PCA
from sklearn.linear_model import LogisticRegression
from sklearn.externals import joblib 

# 算法部分
# 加载数据集
def loadDataSet(k):  # k代表在10张图片中选择几张作为训练集
    dataSetDir = 'C:/FaceDB_orl'
    # 显示文件夹内容
    choose = random.permutation(10) + 1  # 随机排序1-10 (0-9）+1
    train_face = zeros((40 * k, 112 * 92))
    train_face_number = zeros(40 * k)
    test_face = zeros((40 * (10 - k), 112 * 92))
    test_face_number = zeros(40 * (10 - k))
    for i in range(40):  # 40个人
        people_num = i + 1
        for j in range(10):  # 每个人有10个不同的脸
            if j < k:  # 测试集
                filename = dataSetDir + '/' + str(people_num) + '/0' + str(choose[j]) + '.png'
                img = img2vector(filename)
                train_face[i * k + j, :] = img
                train_face_number[i * k + j] = people_num
            else:
                filename = dataSetDir + '/' + str(people_num) + '/0' + str(choose[j]) + '.png'
                img = img2vector(filename)
                test_face[i * (10 - k) + (j - k), :] = img
                test_face_number[i * (10 - k) + (j - k)] = people_num
    return train_face, train_face_number, test_face, test_face_number


# 将图片转换成矩阵
def img2vector(filename):
    img = cv2.imread(filename, 0)  # 读入灰度值
    rows, cols = img.shape
    imgVector = zeros((1, rows * cols))
    imgVector = reshape(img, (1, rows * cols))  # 将2维转成1维
    return imgVector


def facefind():
    # 获取训练集
    train_face, train_face_number, test_face, test_face_number = loadDataSet(3)
    # PCA训练训练集，用pca将数据降到30维
    pca = PCA(n_components=30).fit(train_face)
    # 返回测试集和训练集降维后的数据集
    x_train_pca = pca.transform(train_face)
    x_test_pca = pca.transform(test_face)
    # 逻辑回归训练
    classirfier = LogisticRegression()
    lr = classirfier.fit(x_train_pca,train_face_number)
    
    #保存模型
    joblib.dump(lr,'lr.model')
    # 计算精确度和召回率
    accuray = classirfier.score(x_test_pca, test_face_number)
    recall = accuray*0.7

    return accuray,recall,pca

#界面部分
def choosepic():  # 选择图片函数
    file_path = filedialog.askopenfilename()  # 加载文件
    path.set(file_path)
    img_open = Image.open(file.get())
    img = ImageTk.PhotoImage(img_open)
    pic_label.config(image=img)
    pic_label.image = img

    string = str(file.get())
    
    #预测的人
    predict = img2vector(string)
    #加载模型
    LR = joblib.load('lr.model')
    predict_people = LR.predict(pca.transform(predict))

    string1 = str("编号：%s 精确度：%f 召回率：%f"%(predict_people,accuray,recall))
    showinfo(title='图像分析', message = string1)

    
#初始化Tk（）
accuray,recall,pca = facefind()
root = Tk()#root便是你布局的根节点了，以后的布局都在它之上
root.geometry('260x140')
root.title("人脸识别系统")#设置窗口标题
root.resizable(width=False,height=False)#设置窗口是否可变
root.tk.eval('package require Tix')#引入升级包，这样才能使用升级的组合控件
path = StringVar()#跟踪变量的值的变化

Button(root,text='选择图片',command=choosepic,width=1,height=1).grid(row=1,column=1,sticky=W+E+N+S,padx=40,pady=20)#command指定其回调函数
file = Entry(root,state='readonly',text=path)
file.grid(row=0,column=1,sticky=W+E+S+N,padx=6,pady=20)#用作文本输入用

pic_label = Label(root,text='图片',padx=30,pady=10)
pic_label.grid(row=0,column=2,rowspan=4,sticky=W+E+N+S)
root.mainloop()

**

结果展示：