When will you grow up?

Haar Cascade Classifiers를 이용한 얼굴 인식 본문

02. Study/Computer Vision(openframworks&opencv)

Haar Cascade Classifiers를 이용한 얼굴 인식

미카이 2019. 10. 7. 01:10

이번에는 Haar Cascade를 이용하여 얼굴인식을 해보자.

최근들어 yolo, facenet, vgg-face등 다양한 딥러닝을 이용하여 많이 인식을 하고있는데 낭중에 기회가 되면 포스팅을 해야겠다.

 

약 10줄되는 코드로 학습된 cascade xml를 불러와서 인식을 해보자.

사용되는 python버전은 3.6, opencv 3.4.7를 이용하여 opencv다운은 여기서 할 수 있다.

학습을 위한 데이터도 충분하지 않아, 사전학습된 xml를 이용하여 검출해보자. xml은 opencv git에서 받을 수 있다. 다운

 

원본
haar_cascade classifiers 적용

전체코드

 

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

# load packages

import cv2

import numpy as np

 

# 테스트 이미지 불러오기

image = cv2.imread("images.jpg")

 

# RGB -> Gray로 변환

# 얼굴 찾기 위해 그레이스케일로 학습되어 있기때문에 맞춰줘야 한다.

image_gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

 

# 정면 얼굴 인식용 cascade xml 불러오기

# 그 외에도 다양한 학습된 xml이 있으니 테스트해보면 좋을듯..

face_cascade = cv2.CascadeClassifier('haarcascades/haarcascade_frontalface_default.xml')

 

# 이미지내에서 얼굴 검출

faces = face_cascade.detectMultiScale(image_gray, 1.35)

 

# 얼굴 검출되었다면 좌표 정보를 리턴받는데, 없으면 오류를 뿜을 수 있음. 

for (x,y,w,h) in faces:

    cv2.rectangle(image,(x,y),(x+w,y+h),(255,0,0),2# 원본 영상에 위치 표시

    roi_gray = image_gray[y:y+h, x:x+w] # roi 생성

    roi_color = image[y:y+h, x:x+w] # roi 

cv2.imshow('img',image)

cv2.waitKey(0)

cv2.destroyAllWindows()

Colored by Color Scripter

cs

 

ref: https://docs.opencv.org/4.1.0/d7/d8b/tutorial_py_face_detection.html

 

OpenCV: Face Detection using Haar Cascades

Goal In this session, We will see the basics of face detection using Haar Feature-based Cascade Classifiers We will extend the same for eye detection etc. Basics Object Detection using Haar feature-based cascade classifiers is an effective object detection

docs.opencv.org

뭔가 학습하고 원리를 이해해보고 싶다면,

https://darkpgmr.tistory.com/70

 

OpenCV Haar/cascade training 튜토리얼

OpenCV의 Haar classifier, Cascade classifier를 학습시키기 위한 샘플 데이터 생성법 및 training 방법에 대한 상세 메뉴얼입니다. Haar training이나 cascade training에 대한 내용은 OpenCV 웹 매뉴얼이나 Nao..

darkpgmr.tistory.com

 

Comments