HowOldRobot 是微软推出的一个小服务,实现识别图片中的人脸以及预测图片中人的年龄,预测模型可能比较复杂,但是想做一个简单的人脸轮廓的识别还是比较简单的,下面记录使用Opencv工具包里的haarcascade
进行识别的过程。
import matplotlib.pyplot as plt
%matplotlib inline
import sys, cv2
imagePath = '3jpg.jpg'
image = cv2.imread(imagePath)
# plt.imshow(image)
# plt.imshow(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))
faceCascade = cv2.CascadeClassifier('F:\BaiduYunDownload\opencv\sources\data\haarcascades\haarcascade_frontalface_default.xml')
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# plt.imshow(gray)
faces = faceCascade.detectMultiScale(
gray,
scaleFactor=1.1,
minNeighbors=5,
minSize=(30,30),
flags=cv2.CASCADE_SCALE_IMAGE)
print faces
[[401 26 91 91]
[ 47 26 108 108]
[238 44 60 60]]
import random
font = cv2.FONT_HERSHEY_SIMPLEX
for (x, y, w, h) in faces:
cv2.rectangle(image, (x, y), (x+w, y+h), (14, 201, 255), 2)
cv2.putText(image, str(random.randrange(20, 30)), (x+(w/2)-18, y-10), font, 1,(14, 201, 255), 2)
plt.imshow(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))