0

I wrote the script by looking at this website and it works perfectly but the only problem when I run it on my computer is it only detects the eye region. https://pythonprogramming.net/haar-cascade-face-eye-detection-python-opencv-tutorial/

Below is the script I wrote based on the website.

import numpy as np
import cv2

face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml')
eye_cascade = cv2.CascadeClassifier('haarcascade_eye.xml')

image = cv2.imread('frame119.jpg')  
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray, 1.3, 5)

for (x,y,w,h) in faces:
    cv2.rectangle(image,(x,y),(x+w,y+h),(255,0,0),2)
    roi_gray = gray[y:y+h, x:x+w]
    roi_color = image[y:y+h, x:x+w]

    eyes = eye_cascade.detectMultiScale(roi_gray)
    for (ex,ey,ew,eh) in eyes:
        cv2.rectangle(roi_color,(ex,ey),(ex+ew,ey+eh),(0,255,0),2)

cv2.imshow('image', image)
cv2.waitKey(0)
cv2.destroyAllWindows()

do I need to add an additional line to fix the problem? Also, the image is in 680x480 dimension, I think that maybe one of the reason why it only detects eye region of the image but I do not have any idea regarding that.

Thank you for the help.

Ling
  • 891
  • 5
  • 17
  • 40

1 Answers1

1

It is not possible to detect eyes if no face is detected .

so try these modification 

import numpy as np
import cv2

face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml')
eye_cascade = cv2.CascadeClassifier('haarcascade_eye.xml')

image = cv2.imread('frame119.jpg')  
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray, 1.3, 5)

print len(faces)   # it will print no of faces detected 
for (x,y,w,h) in faces:
    cv2.rectangle(image,(x,y),(x+w,y+h),(255,0,0),2)
    roi_gray = gray[y:y+h, x:x+w]
    roi_color = image[y:y+h, x:x+w]
    cv2.imshow('face',roi_color)  # It will show a cropped face , if face is detected
    cv2.waitKey()    
    eyes = eye_cascade.detectMultiScale(roi_gray)
    for (ex,ey,ew,eh) in eyes:
        cv2.rectangle(roi_color,(ex,ey),(ex+ew,ey+eh),(0,255,0),2)

cv2.imshow('image', image)
cv2.waitKey(0)
cv2.destroyAllWindows()
Shahrukh khan
  • 372
  • 2
  • 9
  • I did run the modified script that you give but for the cropped face part, it only shows eyes and the number of faces detected shows 2 which is both eyes.I want to make the program to detect the whole face and the eye region just like in the example from the website but when I run the program, both rectangle detect the eye region. I guess this must be problem with the webcam maybe? Thank you for your help. – Ling Aug 19 '16 at 11:53
  • the probem may be in the .xml file , plz verify if it is the valid file for face detection and it is not renamed . – Shahrukh khan Aug 19 '16 at 16:02
  • the probem may be in the .xml file , plz verify if it is the valid file for face detection and it is not renamed ,try with some other image , try with some other face cascade classifier like "haarcascade_frontalface_alt_tree.xml" – Shahrukh khan Aug 19 '16 at 16:14
  • thank you for your suggestion. I used haarcascade_frontalface_alt_tree.xml like you suggest. even the rectangle problem vanished but it did not detect all faces in the picture. – Ling Aug 20 '16 at 01:43
  • i solved it. using haarcascade_frontalface_alt.xml. but there is some problem where it did not detect people with glasses. there might be other cases for that. nevertheless the program runs perfectly. the rectangle pointed out the correct spot. thank you for your help. – Ling Aug 20 '16 at 01:51