Face recognition using OpenCV (Haar-like feature classifier)

Haar-like feature classifier

"Cascade of boosted classifiers using Haar-like features" is a method called Haar-like that uses a rectangular pattern consisting of light and dark for detection. The Haar-like feature and the target image are overlaid to determine if there is a pattern. 　

The following Haar-like feature classifiers are provided in OpenCV in advance. [OpenCV Install Dir]/Library/etc/haarcascades/

This time, I will try face recognition using OpenCV's Haar-like feature classifier.

OpenCV OpenCV (Open Source Computer Vision Library) is a collection of BSD-licensed video / image processing libraries. There are many algorithms for image filtering, template matching, object recognition, video analysis, machine learning, and more.

■ Example of motion tracking using OpenCV (OpenCV Google Summer of Code 2015) https://www.youtube.com/watch?v=OUbUFn71S4s

■ Click here for installation and easy usage Install OpenCV 3 (core + contrib) in Python 3 environment & Difference between OpenCV 2 and OpenCV 3 & simple operation check ★ Please install core + opencv_contrib to run the motion template.

■ Click here for still image filtering Try edge detection with OpenCV Perform various filters with OpenCV (gradient, high-pass, Laplacian, Gaussian) Extract feature points with OpenCV (AgastFeature, FAST, GFTT, MSER, AKAZE, BRISK, KAZE, ORB, SimpleBlob)

■ Click here for processing video files Try converting videos in real time with OpenCV Try converting webcam / camcorder video in real time with OpenCV Draw optical flow in real time with OpenCV (Shi-Tomasi method, Lucas-Kanade method) Object tracking using OpenCV (tracking feature points specified by mouse by Lucas-Kanade method Motion template analysis using OpenCV (recognizes objects and their moving directions in real time)

program

The flow is as follows.

1. Create a Haar-like feature classifier with cv2.CascadeClassifier ([Haar-like feature file.xml])
2. Specify an image for the Haar-like feature classifier
3. Enclose the detected area in the image with a rectangle

face.py

import cv2

# Haar-like loading feature classifier
face_cascade = cv2.CascadeClassifier('haarcascades/haarcascade_frontalface_default.xml')
eye_cascade = cv2.CascadeClassifier('haarcascades/haarcascade_eye.xml')

#Loading image file
img = cv2.imread('face.png')

#Grayscale conversion
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

#Face detection
faces = face_cascade.detectMultiScale(gray)
for (x,y,w,h) in faces:
    #Enclose the detected face in a rectangle
    cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2)
    #Face image (grayscale)
    roi_gray = gray[y:y+h, x:x+w]
    #Face g increase (color scale)
    roi_color = img[y:y+h, x:x+w]
    #Detect eyes from inside the face
    eyes = eye_cascade.detectMultiScale(roi_gray)
    for (ex,ey,ew,eh) in eyes:
        #Enclose the detected eye in a rectangle
        cv2.rectangle(roi_color,(ex,ey),(ex+ew,ey+eh),(0,255,0),2)

#Image display
cv2.imshow('img',img)

#Press any key to finish
cv2.waitKey(0)
cv2.destroyAllWindows()

Execution result

If it is such an image, it will be recognized like this. I recognize all the faces, but the eyes don't recognize it very much, probably because of the vague image.

If it is such an image, it will be recognized like this. The whole face is quite recognizable. On the other hand, the eyes recognize if the line of sight is facing the front.

If you understand the detection patterns of your eyes and profile to some extent, you will be able to guess what you are looking at and what is ahead of you.