[PYTHON/OPENCV] CascadeClassifier 클래스 : detectMultiScale 메소드를 사용해 Haar 직렬형 분류기 사용하기

■ CascadeClassifier 클래스의 detectMultiScale 메소드를 사용해 Haar 직렬형 분류기를 만드는 방법을 보여준다.

▶ 예제 코드 (PY)


import cv2

faceCascadeClassifier = cv2.CascadeClassifier("haarcascade_frontalface_default.xml")
eyeCascadeClassifier  = cv2.CascadeClassifier("haarcascade_eye.xml")

imageNDArray          = cv2.imread("source.jpg")
grayscaleImageNDArray = cv2.cvtColor(imageNDArray, cv2.COLOR_BGR2GRAY)

faceNDArray = faceCascadeClassifier.detectMultiScale(grayscaleImageNDArray, 1.3, 5)

for (faceX, faceY, faceWidth, faceHeight) in faceNDArray:
    left   = faceX
    top    = faceY
    right  = faceX + faceWidth
    bottom = faceY + faceHeight

    cv2.rectangle(imageNDArray, (left, top), (right, bottom), (255, 0, 0), 2)

    faceImageNDArray          = imageNDArray[top:bottom, left:right]
    faceGrayscaleImageNDArray = grayscaleImageNDArray[top:bottom, left:right]

    eyeNDArray = eyeCascadeClassifier.detectMultiScale(faceGrayscaleImageNDArray)

    for(eyeX, eyeY, eyeWidth, eyeHeight) in eyeNDArray:
        cv2.rectangle(faceImageNDArray, (eyeX, eyeY), (eyeX + eyeWidth, eyeY + eyeHeight), (0, 255, 0), 2)

cv2.imwrite("target.jpg", imageNDArray)

cv2.imshow("image", imageNDArray)

cv2.waitKey(0)

cv2.destroyAllWindows()

import cv2

faceCascadeClassifier = cv2.CascadeClassifier("haarcascade_frontalface_default.xml")

eyeCascadeClassifier = cv2.CascadeClassifier("haarcascade_eye.xml")

imageNDArray = cv2.imread("source.jpg")

grayscaleImageNDArray = cv2.cvtColor(imageNDArray, cv2.COLOR_BGR2GRAY)

faceNDArray = faceCascadeClassifier.detectMultiScale(grayscaleImageNDArray, 1.3, 5)

for (faceX, faceY, faceWidth, faceHeight) in faceNDArray:

left = faceX

top = faceY

right = faceX + faceWidth

bottom = faceY + faceHeight

cv2.rectangle(imageNDArray, (left, top), (right, bottom), (255, 0, 0), 2)

faceImageNDArray = imageNDArray[top:bottom, left:right]

faceGrayscaleImageNDArray = grayscaleImageNDArray[top:bottom, left:right]

eyeNDArray = eyeCascadeClassifier.detectMultiScale(faceGrayscaleImageNDArray)

for(eyeX, eyeY, eyeWidth, eyeHeight) in eyeNDArray:

cv2.rectangle(faceImageNDArray, (eyeX, eyeY), (eyeX + eyeWidth, eyeY + eyeHeight), (0, 255, 0), 2)

cv2.imwrite("target.jpg", imageNDArray)

cv2.imshow("image", imageNDArray)

cv2.waitKey(0)

cv2.destroyAllWindows()

data.zip