一、人脸模型训练
import os
import cv2 as cv
import sys
import numpy as np
from PIL import Image
def getImageAndLable(path):
faces = []
ids = []
imagePaths = [os.path.join(path, file) for file in os.listdir(path)]
#
face_detector = cv.CascadeClassifier(
'D:\\SoftWareHouse\\opencv\\sources\\data\\haarcascades\\haarcascade_frontalface_default.xml')
for imagePath in imagePaths:
PIL_image = Image.open(imagePath).convert("L")
Np_image = np.array(PIL_image, 'uint8')
face = face_detector.detectMultiScale(Np_image, scaleFactor=1.2, minNeighbors=5)
for x, y, w, h in face:
faces.append(Np_image[y:y + h, x:x + w])
ids.append(int(os.path.split(imagePath)[1].split('.')[0]))
print(ids)
return faces, ids
if __name__ == '__main__':
# 训练图片路径
path = './data/jm'
# 获取图像数组和ID数组
faces, ids = getImageAndLable(path)
# 获取循环对象
recognizer = cv.face.LBPHFaceRecognizer_create()
recognizer.train(faces, np.array(ids))
# 写入训练模型
recognizer.write('trainer/train_demo.yml')
二、在人脸模型中识别人脸
import cv2 as cv
# 加载训练好的模型文件
recognizer = cv.face.LBPHFaceRecognizer_create()
recognizer.read('trainer/train_demo.yml')
# 加载图片
image=cv.imread('face5.jpg')
gray = cv.cvtColor(image, cv.COLOR_BGR2GRAY)
# 加载图片的特征值
face_detector = cv.CascadeClassifier(
'D:\\SoftWareHouse\\opencv\\sources\\data\\haarcascades\\haarcascade_frontalface_default.xml')
faces = face_detector.detectMultiScale(gray)
for x, y, w, h in faces:
cv.rectangle(image, (x, y), (x + w, y + h), color=(255, 0, 0))
# 在训练模型中,找到对应的ID和置信评分,置信评分为0最接近
id, confidence = recognizer.predict(gray[y:y + h, x:x + w])
# 打印ID和置信评分
print('ID:', id, '置信评分:', confidence)
cv.imshow('image', image)
cv.waitKey(0)
cv.destroyAllWindows()