I have an algorithm Viola-Jones in Python. I'm using haarcascade xml, which I load from openCV root file. But there wasn't any xml file for mouth and nose in openCV, so I downloaded these files from EmguCV. Result for detection of face is OK, but detection of eye isn't good and nose with mouth is very bad. I tried to change parameters in face_cascade.detectMultiScale, but it didn't help at all.
My code:
import cv2
import sys
def facedet(img):
face_cascade = cv2.CascadeClassifier('/home/kattynka/opencv/data/haarcascades/haarcascade_frontalface_default.xml')
eye_cascade = cv2.CascadeClassifier('/home/kattynka/opencv/data/haarcascades/haarcascade_eye.xml')
mouth_cascade = cv2.CascadeClassifier('/home/kattynka/opencv/data/haarcascades/haarcascade_mcs_mouth.xml')
nose_cascade = cv2.CascadeClassifier('/home/kattynka/opencv/data/haarcascades/haarcascade_mcs_nose.xml')
img = cv2.imread(img)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray, 1.3, 5)
for (x,y,w,h) in faces:
cv2.rectangle(img, (x,y), (x+w,y+h), (255,0,0), 2)
roi_gray = gray[y:y+h, x:x+w]
roi_color = img[y:y+h, x:x+w]
eyes = eye_cascade.detectMultiScale(roi_gray)
nose = nose_cascade.detectMultiScale(roi_gray)
mouth = mouth_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)
for (nx, ny, nw, nh) in nose:
cv2.rectangle(roi_color, (nx, ny), (nx + nw, ny + nh), (0, 0, 255), 2)
for (mx, my, mw, mh) in mouth:
cv2.rectangle(roi_color, (mx, my), (mx + mw, my + mh), (0, 0, 0), 2)
cv2.namedWindow('image', cv2.WINDOW_NORMAL)
cv2.imshow('image',img)
cv2.waitKey(0)
cv2.destroyAllWindows()
if __name__ == '__main__':
#img = sys.argv[1]
facedet(img)
My question
What am I doing wrong? Is there any simple solution, which will give me a better result?
Output:
Haar cascades perform alright for faces but not so well for smaller individual parts. A better solution is to detect all the face landmarks together. A good algorithm for that is "One Millisecond Face Alignment with an Ensemble of Regression Trees by Vahid Kazemi and Josephine Sullivan, CVPR 2014" which is implemented in Dlib (http://dlib.net/face_landmark_detection.py.html).
This works really well for me.
I found that if you divide the face into 2 sections and have the eyes look for eyes in the top section, and the mouth in the lower section it works really well.
face
--------
| eyes |
|------|
|mouth |
--------
This is a rough illustration of what I did with the code below.
I am aware the the cascade i use is smile, but the mouth doesn't seem to work.
import cv2
import sys
mouthCascade = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_smile.xml')
faceCascade = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_frontalface_default.xml')
eyeCascade = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_eye.xml')
video_capture = cv2.VideoCapture(0)
while True:
# Capture frame-by-frame
ret, frame = video_capture.read()
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
mouth = mouthCascade.detectMultiScale(gray, 1.3, 5)
faces = faceCascade.detectMultiScale(
gray,
scaleFactor=1.1,
minNeighbors=5,
minSize=(30, 30)
)
# Draw a rectangle around the faces
for (x, y, w, h) in faces:
cv2.rectangle(frame, (x, y), (x+w, y+h), (255, 0, 0), 2)
# Draw a rectangle around the faces
roi_gray_mouth = gray[y+(int(h/2)):y+h, x:x+w]
roi_color_mouth = frame[y+(int(h/2)):y+h, x:x+w]
roi_gray_eye = gray[y-(int(h/2)):y+h, x:x+w]
roi_color_eye = frame[y-(int(h/2)):y+h, x:x+w]
mouth = mouthCascade.detectMultiScale(roi_gray_mouth)
eyes = eyeCascade.detectMultiScale(roi_gray_eye)
for (ex,ey,ew,eh) in mouth:
cv2.rectangle(roi_color_mouth, (ex, ey), (ex+ew, ey+eh), (0, 255, 0), 2)
for (eex,eey,eew,eeh) in eyes:
d = int(eew / 2)
cv2.circle(roi_color_eye, (int(eex + eew / 4) + int(d / 2), int(eey + eeh / 4) + int(d / 2)), int(d) ,(0,0,255),2)
# Display the resulting frame
cv2.imshow('Video', frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
# When everything is done, release the capture
video_capture.release()
cv2.destroyAllWindows()
import cv2
import sys
face_cascade = cv2.CascadeClassifier('/home/kattynka/opencv/data/haarcascades/haarcascade_frontalface_default.xml')
eye_cascade = cv2.CascadeClassifier('/home/kattynka/opencv/data/haarcascades/haarcascade_eye.xml')
mouth_cascade = cv2.CascadeClassifier('/home/kattynka/opencv/data/haarcascades/haarcascade_mcs_mouth.xml')
nose_cascade = cv2.CascadeClassifier('/home/kattynka/opencv/data/haarcascades/haarcascade_mcs_nose.xml')
img = cv2.imread(img)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray, 1.3, 5)
for (x,y,w,h) in faces:
cv2.rectangle(img, (x,y), (x+w,y+h), (255,0,0), 2)
roi_gray = gray[y:y+h, x:x+w]
roi_color = img[y:y+h, x:x+w]
eyes = eye_cascade.detectMultiScale(gray, 1.3, 5)
nose = nose_cascade.detectMultiScale(gray, 1.3, 5)
mouth = mouth_cascade.detectMultiScale(gray, 1.7, 11)
for (ex,ey,ew,eh) in eyes:
cv2.rectangle(img, (ex,ey), (ex+ew, ey+eh), (0,255,0), 2)
for (nx, ny, nw, nh) in nose:
cv2.rectangle(img, (nx, ny), (nx + nw, ny + nh), (0, 0, 255), 2)
for (mx, my, mw, mh) in mouth:
cv2.rectangle(img, (mx, my), (mx + mw, my + mh), (0, 0, 0), 2)
cv2.namedWindow('image', cv2.WINDOW_NORMAL)
cv2.imshow('image',img)
cv2.waitKey(0)
cv2.destroyAllWindows()
you can try this code. It worked for me.
Related
im using cv2 to try to take a picture here is my code
import numpy as np
import cv2
face_cascade = cv2.CascadeClassifier('cascades/data/haarcascade_frontalface_alt2.xml')
eye_cascade = cv2.CascadeClassifier('cascades/data/haarcascade_eye.xml')
cap = cv2.VideoCapture(0)
while(True):
ret, frame = cap.read()
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray, scaleFactor=1.1, minNeighbors=5)
for (x, y, w, h) in faces:
#print(x,y,w,h)
roi_gray = gray[y:y+h, x:x+w]
roi_color = frame[y:y+h, x:x+w]
img_item = "my-image.png"
cv2.imwrite(img_item, roi_color)
stroke = 2
color = (0, 255, 0)
end_cord_x = x + w
end_cord_y = y + h
ecolor = (50, 50, 50)
cv2.rectangle(frame, (x,y), (end_cord_x, end_cord_y), color, stroke)
eyes = eye_cascade.detectMultiScale(roi_gray)
for (ex, ey, ew, eh) in eyes:
cv2.rectangle(roi_color, (ex,ey), (ex+ew,ey+eh), ecolor, 1)
cv2.imshow('frame',frame)
if cv2.waitKey(20) & 0xFF == ord('q'):
break
cap.release()
cv2.destroyAllWindows()
i get the error:
File "c:NothereForPrivacyReasons\imagerecognition\face\picture.py", line 3, in
face_cascade = cv2.CascadeClassifier('cascades/data/haarcascade_frontalface_alt2.xml')
AttributeError: module 'cv2' has no attribute 'CascadeClassifier'
I have tried to use cv2.putText and it appears to show the position based on the the top right of the window and not the actual center of the image. It will probably be an obvious fix since I just started using opencv
import os
import numpy as np
font = cv2.FONT_HERSHEY_SIMPLEX
org = (50, 50)
fontScale = 1
color = (255, 0, 0)
radius = 3
thickness = 2
cascPath=os.path.dirname(cv2.__file__)+"/data/haarcascade_frontalface_default.xml"
faceCascade = cv2.CascadeClassifier(cascPath)
video_capture = cv2.VideoCapture(0)
while (True):
ret, frames = video_capture.read()
gray = cv2.cvtColor(frames, cv2.COLOR_BGR2GRAY)
faces = faceCascade.detectMultiScale(
gray,
scaleFactor=1.1,
minNeighbors=5,
minSize=(30, 30),
flags=cv2.CASCADE_SCALE_IMAGE
)
for (x, y, w, h) in faces:
cv2.rectangle(frames, (x, y), (x+w, y+h), (0, 255, 0), 2)
text = (x+w//2), (y+h//2)
cv2.circle(frames, (cx, cy), radius, (255, 0, 0), -1)
cv2.putText(frames, str(text), org, font, fontScale, color, thickness)
cv2.imshow('Video', frames)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
video_capture.release()
cv2.destroyAllWindows()
With the cv2.getTextSize() function, you can calculate the pixel size of the text you will put and subtract it from the text's position. In this way, the text will be right on the center.
text = (x+w//2), (y+h//2)
text_size,t = cv2.getTextSize(text=str(text), fontFace=cv2.FONT_HERSHEY_DUPLEX, fontScale=1, thickness=1)
text_size_x,text_size_y = text_size
text_pos = (x+w//2)-(text_size_x//2), (y+h//2)+(text_size_y//2)
Here is a working code
import os
import numpy as np
import cv2
font = cv2.FONT_HERSHEY_SIMPLEX
org = (50, 50)
fontScale = 1
color = (255, 0, 0)
radius = 3
thickness = 2
cascPath=os.path.dirname(cv2.__file__)+"/data/haarcascade_frontalface_default.xml"
faceCascade = cv2.CascadeClassifier(cascPath)
video_capture = cv2.VideoCapture(0)
while (True):
ret, frames = video_capture.read()
gray = cv2.cvtColor(frames, cv2.COLOR_BGR2GRAY)
faces = faceCascade.detectMultiScale(
gray,
scaleFactor=1.1,
minNeighbors=5,
minSize=(30, 30),
flags=cv2.CASCADE_SCALE_IMAGE
)
for (x, y, w, h) in faces:
cv2.rectangle(frames, (x, y), (x+w, y+h), (0, 255, 0), 2)
text = (x+w//2), (y+h//2)
text_size,t = cv2.getTextSize(text=str(text), fontFace=cv2.FONT_HERSHEY_DUPLEX, fontScale=1, thickness=1)
text_size_x,text_size_y = text_size
text_pos = (x+w//2)-(text_size_x//2), (y+h//2)+(text_size_y//2)
#cv2.circle(frames, (cx, cy), radius, (255, 0, 0), -1)
cv2.putText(frames, str(text), text_pos, font, fontScale, color, thickness)
cv2.imshow('Video', frames)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
video_capture.release()
cv2.destroyAllWindows()
My original idea can be seen in the code below but it's still detecting eyes on the lower half of my face. The goal is to make it only scan the upper half of my face, therefore weeding out incorrect matches.
import cv2
face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml')
eye_cascade = cv2.CascadeClassifier('haarcascade_eye.xml')
cap = cv2.VideoCapture(0) # sets up webcam
while 1: # capture frame, converts to greyscale, looks for faces
ret, img = cap.read()
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray, 1.3, 5)
for (x, y, w, h) in faces: # draws box around face
cv2.rectangle(img, (x, y), (x + w, y + h), (255, 0, 0), 2)
roi_gray = gray[y:y + h, x:x + w]
roi_color = img[y:y + h, x:x + w]
half_point = y
print("half point: " + str(half_point))
eyes = eye_cascade.detectMultiScale(roi_gray) # looks for eyes
for (ex, ey, ew, eh) in eyes: # draws boxes around eyes
check_point = ey
print("check_point: " + str(check_point))
if check_point > half_point:
pass
else:
cv2.rectangle(roi_color, (ex, ey), (ex + ew, ey + eh), (0, 255, 0), 2)
cv2.imshow('img', img)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
cap.release()
cv2.destroyAllWindows()
I only modified the line 15 and 16
cv2.rectangle(img, (x, y), (x + w, y + int(h / 2)), (255, 0, 0), 2)
roi_gray = gray[y:y + int(h / 2), x:x + w]
The full code:
import cv2
face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml')
eye_cascade = cv2.CascadeClassifier('haarcascade_eye.xml')
cap = cv2.VideoCapture(0) # sets up webcam
while 1: # capture frame, converts to greyscale, looks for faces
ret, img = cap.read()
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray, 1.3, 5)
for (x, y, w, h) in faces: # draws box around face
cv2.rectangle(img, (x, y), (x + w, y + int(h / 2)), (255, 0, 0), 2) #Modified
roi_gray = gray[y:y + int(h / 2), x:x + w] #Modified
roi_color = img[y:y + h, x:x + w]
half_point = y
print("half point: " + str(half_point))
eyes = eye_cascade.detectMultiScale(roi_gray) # looks for eyes
for (ex, ey, ew, eh) in eyes: # draws boxes around eyes
check_point = ey
print("check_point: " + str(check_point))
if check_point > half_point:
pass
else:
cv2.rectangle(roi_color, (ex, ey), (ex + ew, ey + eh), (0, 255, 0), 2)
cv2.imshow('img', img)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
cap.release()
cv2.destroyAllWindows()
But, I recommend to use dlib insead. It's more reliable.
Here is my example:
import numpy as np
import cv2
import dlib
cap = cv2.VideoCapture(0)
predictor = dlib.shape_predictor("shape_predictor_68_face_landmarks.dat")
detector = dlib.get_frontal_face_detector()
def draw_on_frame(eye):
coordinates = np.array([])
for i in eye:
x = landmarks.part(i).x
y = landmarks.part(i).y
cv2.circle(frame, (x, y), 3, (0, 0, 255), -1)
coordinates = np.append(coordinates, [x, y])
x1, y1, w1, h1 = cv2.boundingRect(coordinates.reshape(-1, 2).astype(int))
cv2.rectangle(frame, (x1, y1), (x1 + w1, y1 + h1), (0, 255, 0), 1)
return x1, y1, w1, h1
while (cap.isOpened()):
ret, frame = cap.read()
if ret == True:
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
faces = detector(gray)
for face in faces:
landmarks = predictor(gray, face)
left_eye = range(36, 42)
right_eye = range(42, 48)
left = draw_on_frame(left_eye)
right = draw_on_frame(right_eye)
roi_left = frame[left[1]:left[1]+left[3], left[0]:left[0]+left[2]]
roi_right = frame[right[1]:right[1] + right[3], right[0]:right[0] + right[2]]
cv2.imshow('frame', frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
else:
break
cap.release()
cv2.destroyAllWindows()
I am getting a NameError: name 'eyes' not found while trying to run an OpenCV project in Python on cmd. I am using Python 2.7 and OpenCV 2.4.13, which I think is not a problem.
import cv2
import numpy as np
face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml')
eye_cascade = cv2.CascadeClassifier('haarcascade_eye.xml')
cap = cv2.VideoCapture(1)
while True:
ret, img = cap.read()
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray, 1.3, 5)
for(x,y,w,h) in faces:
cv2.reactangle(img, (x,y), (x+w,y+h), (255,0,0), 2)
roi_gray = gray[y:y+h, x:x+w]
roi_color = img[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('img',img)
k = cv2.waitKey(30) & 0xff
if k == 27:
break
cap.release()
cv2.destroyAllWindows()
Error:
File "detect.py" , line 19, in
for(ex,ey,ew,eh) in eyes:
NameError: name 'eyes' is not defined
Problem of indentation, just like this one:
eyes is out of scope when you go out of the faces loop.
while True:
ret, img = cap.read()
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray, 1.3, 5)
for(x,y,w,h) in faces:
cv2.reactangle(img, (x,y), (x+w,y+h), (255,0,0), 2)
roi_gray = gray[y:y+h, x:x+w]
roi_color = img[y:y+h, x:x+w]
eyes = eye_cascade.detectMultiScale(roi_gray)
# this one should be inside the 'faces' loop
for(ex,ey,ew,eh) in eyes:
cv2.rectangle(roi_color, (ex, ey), (ex+ew,ey+eh), (0,255,0), 2)
cv2.imshow('img',img)
k = cv2.waitKey(30) & 0xff
if k == 27:
break
I used the following code to detect a face using Haar cascade classifiers provided by OpenCv Python. But the faces are not detected and the square around the face is not drawn. How to solve this?
import cv2
index=raw_input("Enter the index No. : ")
cascPath = "haarcascade_frontalface_default.xml"
faceCascade = cv2.CascadeClassifier(cascPath)
cap = cv2.VideoCapture(0)
cont=0
while(True):
# Capture frame-by-frame
ret, frame = cap.read()
# Our operations on the frame come here
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
faces = faceCascade.detectMultiScale(
gray,
scaleFactor=1.1,
minNeighbors=10,
minSize=(30, 30),
flags = cv2.cv.CV_HAAR_SCALE_IMAGE
)
for (x, y, w, h) in faces:
#cv2.rectangle(frame, (x, y), (x+w, y+h), (0, 255, 0), 2)
cv2.rectangle(frame, (x, y), (x+w, y+h), (0, 255, 0), 2)
# Display the resulting frame
cv2.imshow('frame',frame)
inpt=cv2.waitKey(1)
if inpt & 0xFF == ord('q'):
break
elif inpt & 0xFF == ord('s') :
#name='G:\XCODRA\Integrated_v_01\EigenFaceRecognizer\img2'+index+"."+(str(cont))+".png"
name='IC_image\\'+index+"."+(str(cont))+".png"
resized = cv2.resize(gray,None,fx=200, fy=200, interpolation = cv2.INTER_AREA)
img=cv2.equalizeHist(resized)
cv2.imwrite(name,img)
print cont
cont+=1
Use the full path for the classifier.