Face detection based on OpenCV and dlib+fr and face alignment based on Dlib

In the course of media big data , Senior brothers and sisters introduced face detection and face alignment based on various common databases , And did a series of experiments .

One 、 be based on OpenCV Face detection

1. install OpenCV:

pip install -i https://pypi.tuna.tsinghua.edu.cn/simple opencv-python

2. Get the location of the classifier file ：

see python The path of ：

where python

stay Python Under the path /Lib/sitepackages/cv2/data/haarcascade_frontalface_default.xml' That is, the location of the classifier file

3. Code testing ：

import cv2

filename = "test3.jpg"
### establish ⼀ Cascaded classifiers , load ⼀ individual  .xml⽂ Pieces of 
face_cascade =cv2.CascadeClassifier('D:\Anaconda3\Lib\site-packages\cv2\data\haarcascade_frontalface_default.xml')
### Load image 
img = cv2.imread(filename)
### Convert to grayscale image 
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
### send ⽤ On ⾯ Create a classifier to detect ⼈ face ,faces To recognize the rectangular box vector group of the face ,⼏ Every face has ⼏ Moments   Shape frame , Vector group ⾥ There is ⼏ Group vectors , preservation ⼈ The coordinates of the face and ⼤⼩.
faces = face_cascade.detectMultiScale(gray,1.3,5)
### scaleFactor Express ⼈ The time chart of each iteration in the face detection process ⽚ Compression ratio of （1.3）
### minNeighbors： Every ⼈ The face rectangle retains the number of nearest neighbors ⽬ The most ⼩ value （5）
### Draw a rectangle on the original drawing 
### Traverse faces Rectangular box vector group ⾥ Every time ⼀ Group rectangles （ That is, every ⼀ Open your face ）

for (x, y, w, h) in faces:
    img = cv2.rectangle(img, (x, y), (x + w, y + h), (255, 0, 0), 2)

cv2.imshow('Person Detected!', img)
cv2.waitKey(0)
cv2.destroyAllWindows()

detection result ：

  Two 、 be based on face_recognition Face detection

Face-Recognition The emissary of ⽤ rely on Dlib, So you need to install it first Dlib：

1. Dlib install ：

Install... With the following command shipment ⾏ Environmental Science ：

## The environment that needs to be installed ：python+opencv+dlib+face_recognition( On the first two ⾯ Installed )
1. pip install cmake
2. pip install dlib ( If an error , Refer to this link to ⾏ install ）
https://blog.csdn.net/qq_58691861/article/details/119116148
https://zhuanlan.zhihu.com/p/464846060
3.pip install face_recognition

 2. After installation , Run code ：

import face_recognition
import cv2
### Load image ⽂ Pieces of （.jpg,.png etc. ）, The data returned is Numpy Array , Record the figure ⽚ The eigenvectors of all pixels 
image = face_recognition.load_image_file("test3.jpg")

### Locate all in the diagram ⼈ Pixel position of the face , The return value is in the form of a list , Every... In the list ⼀⾏ yes ⼀ Zhang ⼈ Position information of the face , Include  【top, right, bottom, left】  This is a ⼀ Group tuples .
face_locations_noCNN=face_recognition.face_locations(image)
face_recognition.face_locations(image,model='cnn')


print("face_location_noCNN:")
print(face_locations_noCNN)

face_num2=len(face_locations_noCNN)
print(face_num2)

### Look at the original picture 
org = cv2.imread("test3.jpg")
img = cv2.imread("test3.jpg")
cv2.imshow("test3.jpg",img)
cv2.waitKey()

### Face detection 
for i in range(0,face_num2):
    top = face_locations_noCNN[i][0]
    right = face_locations_noCNN[i][1]
    bottom = face_locations_noCNN[i][2]
    left = face_locations_noCNN[i][3]
    start = (left, top)
    end = (right, bottom)
    color = (0,255,255)
    thickness = 2
    cv2.rectangle(org, start, end, color, thickness)
cv2.imshow("no cnn ",org)
cv2.waitKey(0)
cv2.destroyAllWindows()

detection result ：

  It can be seen that , Compared with OpenCV, be based on face_recognition The accuracy of face detection is higher

3、 ... and 、 Face feature point detection

import cv2
import dlib
path = "test4.jpg"
# Read graph ⽚
img = cv2.imread(path)
# graying 
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# Get the face position detector 


detector = dlib.get_frontal_face_detector()
dets = detector(gray, 1)

predictor =dlib.shape_predictor("D:\Anaconda3\Lib\site-packages\shape_predictor_68_face_landmarks.dat")

for face in dets:
    shape = predictor(img, face) #  seek ⼈ Facial 68 Points 
    for pt in shape.parts():
        pt_pos = (pt.x, pt.y)
        # Parameters ： chart ⽚, round ⼼,  radius ,  Typeface ⾊, The font size ）
        cv2.circle(img, pt_pos, 2, (0, 255, 0), 1)
    cv2.imshow("image", img)
cv2.waitKey(0)
cv2.destroyAllWindows()