使用OpenCV+PCA+KNN/SVM进行人脸检测和识别-Python

墨蓝 2022-05-29 05:47 366阅读 0赞

**注：需要在python中安装opencv库和sklearn库，opencv调用摄像头并检测人脸，sklearn提供PCA、SVM和KNN库。**  
**本文共识别三人，所有人脸图像保存在train文件夹内，并且三人图像分别存至s1、s2和s3文件夹内，每个人100张图。图像进行PCA降维前，全部处理为224\*224大小。**  
**如何创建自己的数据集，可见**[使用OpenCV调用摄像头检测人脸并截图][OpenCV]。  
**OpenCV人脸检测器haarcascade\_frontalface\_alt.xml和本文代码可在[https://github.com/TimeIvyace/OpenCV-PCA-KNN-SVM\_face\_recognition.git][https_github.com_TimeIvyace_OpenCV-PCA-KNN-SVM_face_recognition.git]中下载。**

本文代码先将数据集中所有图像进行PCA降维，然后通过摄像头获取人脸，使用同样类型的PCA降维，最后使用KNN或SVM进行分类。

import numpy as np
    from sklearn.decomposition import PCA
    from sklearn.svm import SVC
    from sklearn.neighbors import KNeighborsClassifier
    import cv2
    from PIL import Image
    
    IMAGE_SIZE = 224
    
    def LoadData(): #载入训练数据集
        data = []
        label = []
        num = 100
        path_cwd = "/train/"
        for j in range(1, 4):
            path = path_cwd + 's' + str(j)
            for number in range(num):
                path_full = path + '/' + str(number) +'.jpg'
                image = Image.open(path_full).convert('L')
                image = image.resize((IMAGE_SIZE, IMAGE_SIZE), Image.ANTIALIAS)
                img = np.reshape(image, (1, IMAGE_SIZE*IMAGE_SIZE))
                data.extend(img)
            label.extend(np.ones(num, dtype=np.int) * j)
        data = np.reshape(data, (num*j, IMAGE_SIZE*IMAGE_SIZE))
        return np.matrix(data), np.matrix(label).T       #返回数据和标签
    
    
    def svm(trainDataSimplified, trainLabel, testDataSimplified):
        clf3 = SVC(C=3.0)  # C为分类数目
        clf3.fit(trainDataSimplified, trainLabel)
        return clf3.predict(testDataSimplified)
    
    def knn(neighbor, traindata, trainlabel, testdata):
        neigh = KNeighborsClassifier(n_neighbors=neighbor)
        neigh.fit(traindata, trainlabel)
        return neigh.predict(testdata)
    
    if __name__ == '__main__':
    
        Data, Label = LoadData()
        pca = PCA(0.9, True, True)  # 建立pca类，设置参数，保留90%的数据方差
        trainDataS = pca.fit_transform(Data)  # 拟合并降维训练数据
    
        #框住人脸的矩形边框颜色
        color = (0, 255, 0)
    
        #捕获指定摄像头的实时视频流
        cap = cv2.VideoCapture(0)
    
        #人脸识别分类器本地存储路径
        cascade_path = "haarcascade_frontalface_alt.xml"
    
        #循环检测识别人脸
        while True:
            _, frame = cap.read()   #读取一帧视频
    
            #图像灰化，降低计算复杂度
            frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    
            #使用人脸识别分类器，读入分类器
            cascade = cv2.CascadeClassifier(cascade_path)
    
            #利用分类器识别出哪个区域为人脸
            faceRects = cascade.detectMultiScale(frame_gray, scaleFactor = 1.2, minNeighbors = 3, minSize = (32, 32))
            if len(faceRects) > 0:
                for faceRect in faceRects:
                    x, y, w, h = faceRect
    
                    #截取脸部图像提交给模型识别这是谁
                    m = frame_gray[y - 10: y + h + 10, x - 10: x + w + 10]
    
                    top, bottom, left, right = (0, 0, 0, 0)
                    image = m
                    # 获取图像尺寸
                    h, w = image.shape
    
                    # 对于长宽不相等的图片，找到最长的一边
                    longest_edge = max(h, w)
    
                    # 计算短边需要增加多上像素宽度使其与长边等长
                    if h < longest_edge:
                        dh = longest_edge - h
                        top = dh // 2
                        bottom = dh - top
                    elif w < longest_edge:
                        dw = longest_edge - w
                        left = dw // 2
                        right = dw - left
                    else:
                        pass
    
                    BLACK = [0]
    
                    # 给图像增加边界，是图片长、宽等长，cv2.BORDER_CONSTANT指定边界颜色由value指定
                    constant = cv2.copyMakeBorder(image, top, bottom, left, right, cv2.BORDER_CONSTANT, value=BLACK)
    
                    # 调整图像大小并返回
                    image = cv2.resize(constant, (IMAGE_SIZE, IMAGE_SIZE))
                    img_test = np.reshape(image, (1, IMAGE_SIZE * IMAGE_SIZE))
                    testDataS = pca.transform(img_test)  # 降维测试数据
                    result = svm(trainDataS, Label, testDataS)  # 使用SVM进行分类
                    # result = knn(5, trainDataS, Label, testDataS)  # 使用KNN进行分类，5为最近邻居数
                    faceID = result[0]
    
    
                    #如果是“我”
                    if faceID == 1:
                        cv2.rectangle(frame, (x - 10, y - 10), (x + w + 10, y + h + 10), color, thickness = 2)
    
                        #文字提示是谁
                        cv2.putText(frame,'A',
                                    (x + 30, y + 30),                      #坐标
                                    cv2.FONT_HERSHEY_SIMPLEX,              #字体
                                    1,                                     #字号
                                    (255,0,255),                           #颜色
                                    2)                                     #字的线宽
                    elif faceID == 2:
                        cv2.rectangle(frame, (x - 10, y - 10), (x + w + 10, y + h + 10), color, thickness = 2)
    
                        #文字提示是谁
                        cv2.putText(frame,'B',
                                    (x + 30, y + 30),                      #坐标
                                    cv2.FONT_HERSHEY_SIMPLEX,              #字体
                                    1,                                     #字号
                                    (255,0,255),                           #颜色
                                    2)
                    elif faceID == 3:
                        cv2.rectangle(frame, (x - 10, y - 10), (x + w + 10, y + h + 10), color, thickness = 2)
    
                        #文字提示是谁
                        cv2.putText(frame,'C',
                                    (x + 30, y + 30),                      #坐标
                                    cv2.FONT_HERSHEY_SIMPLEX,              #字体
                                    1,                                     #字号
                                    (255,0,255),                           #颜色
                                    2)
    
            cv2.imshow("find me", frame)
    
            #等待10毫秒看是否有按键输入
            k = cv2.waitKey(10)
            #如果输入q则退出循环
            if k & 0xFF == ord('q'):
                break
    
        #释放摄像头并销毁所有窗口
        cap.release()
        cv2.destroyAllWindows()

[OpenCV]: http://blog.csdn.net/gaoyueace/article/details/79600698
[https_github.com_TimeIvyace_OpenCV-PCA-KNN-SVM_face_recognition.git]: https://github.com/TimeIvyace/OpenCV-PCA-KNN-SVM_face_recognition.git