informramiz / opencv-face-recognition-python Goto Github PK

AttributeError Traceback (most recent call last)
in ()
9 #perform a prediction
10 predicted_img1 = predict(test_img1)
---> 11 predicted_img2 = predict(test_img2)
12 print("Prediction complete")
13

in predict(test_img)
4 def predict(test_img):
5 #make a copy of the image as we don't want to chang original image
----> 6 img = test_img.copy()
7 #detect face from the image
8 face, rect = detect_face(img)

hi Ramiz,

First of all, great tutorial, thank you.

When i (pip) installed the requirement for this tutorial, I've got OpenCV 3.3.0 and it had some issues with your code.

I've done few adjustments to make it work:

1. Changed the creation of the recognizer from
   face_recognizer = cv2.face.createLBPHFaceRecognizer()
   to
   face_recognizer = cv2.face.LBPHFaceRecognizer_create()

2. In the predict function changed this
   label = face_recognizer.predict(face)
   to
   label, confidence = face_recognizer.predict(face)
   The "predict" returns you a value of confidence also, and that addition fixed the TypeError.

OpenCV (3.4.1) Error: Assertion Failed <ssize.width> 0 && ssize.height > 0> in cv::resize, file resize.cpp, line 4044

OpenCV-Face-Recognition-Python.py, Line 223, in faces, labels = prepare_training_data("training-data")

File "OpenCV-Face-Recognition-Python.py, Line 178, in prepare_training_data cv2.imshow("Training on image...", cv2.resize<image, (400,500)))

cv2.imshow("Training on image...", cv2.resize(image, (400, 500)))
cv2.error: OpenCV(3.4.1) resize.cpp:4044: error <215> ssize.width > 0 && ssize height > 0 in function cv::resize

Please is there a solution to this?

1. Assuming some of the faces in the image is tilted, does OpenCV extracts faces and un-rotate them? (i.e. giving coordinates for rectangles where the frame of the rectangle are not parallel to the frame of the image)
2. Can OpenCV detect faces when the face is partially covered (glasses, hygienic masks)?

matrix.cpp:310: error: (-215) s >= 0 in function cv::setSize
i am getting this error at:
-> face_recognizer.train(faces, np.array(labels))

why?

this is a labels np.array:

[0 0 0 1 2 2 2 3 3 3 3 4 4 4 5 5]

and this is a list of faces:

[array([ 36, 22, 244, 244]), array([ 20, 23, 249, 249]), array([ 32, 20, 245, 245]), array([ 39, 26, 239, 239]), array([ 19, 23, 248, 248]), array([ 28, 22, 235, 235]), array([ 28, 11, 243, 243]), array([ 39, 26, 238, 238]), array([ 27, 26, 243, 243]), array([ 21, 21, 257, 257]), array([ 35, 30, 245, 245]), array([ 33, 28, 246, 246]), array([ 24, 20, 257, 257]), array([ 32, 27, 250, 250]), array([ 35, 31, 232, 232]), array([ 35, 25, 236, 236])]

what is a problem?

AttributeError: 'module' object has no attribute 'face'
on line 246 ?

sir ,
You have written a nice code .
I have a doubt -

In the attached pic , face cannot be found .

Please suggest .
Thanks.

Hello!

Ive changed all dir paths to local ones because I want to be able to run the script only by terminal. However when I run it I get the following errors:

File "/home/kokot/Downloads/opencv-face-recognition-python-master/OpenCV-Face-Recognition-Python.py", line 326, in
predicted_img1 = predict(test_img1)
File "/home/kokot/Downloads/opencv-face-recognition-python-master/OpenCV-Face-Recognition-Python.py", line 298, in predict
img = test_img.copy()
AttributeError: 'NoneType' object has no attribute 'copy'

I am getting this error while training my classifier -
face_recog.train(faces,np.array(faceID))

The full code is as follow:

[import cv2
import os
import numpy as np
%config IPCompleter.greedy=True

def facedetection(test_img):

gray_img = cv2.cvtColor(test_img,cv2.COLOR_BGR2GRAY)

face_cascade= cv2.CascadeClassifier("C:\\Users\\Abhijeet Debadwar\\AppData\\Local\\Programs\\Python\\Python38\\Lib\\site-packages\\cv2\\\data\\haarcascade_frontalface_default.xml")

faces = face_cascade.detectMultiScale(gray_img ,1.42,10)

return faces, gray_img

def labels_for_training_data(directory):

faces =[]
faceID= []

for maindirectory, subdirectory,filenames in os.walk(directory):
    
    for filename in filenames:
        
        if filename[0]==".":
            print("skipped")
            continue
            
        id = os.path.basename(maindirectory)
        
        img_path = os.path.join(maindirectory,filename)
        
        print("img_path ",img_path)
        print("image ID: ", id)
        
        test_img = cv2.imread(img_path)
        
        if test_img is None:
            print("IMAGE NOT LOADED PROPERLY")
            continue
                 
        faces_rect, gray_img = facedetection(test_img)
        
        if len(faces_rect)!=1:
            continue
            
        (x,y,w,h)= faces_rect[0]
        
        ROF_gray = gray_img[x:x+w,y:y+h]
        
        faces.append(ROF_gray)
        faceID.append(int(id))
        
        
return faces, faceID

def train_classifier(faces,faceID):

face_recog = cv2.face.LBPHFaceRecognizer_create()

face_recog.train(faces,np.array(faceID))

return face_recog

def draw_rect(test_img,faces):

(x,y,w,h) = faces

cv2.rectange(test_img,(x,y), (x+w,y+h),(255,0,0), 5)

def put_text(test_img,text,x,y):

cv2.putText(test_img,text, (x,y), cv2.FONT_HERSHEY_DUPLEX, 5, (0,255,0),5)

test_img = cv2.imread("D:\testimages\1\IMG_20190611_185155.jpg",1)

face_detected, gray_img = facedetection(test_img)

faces, faceID = labels_for_training_data("D:\testimages")

face_recog = train_classifier( faces, faceID)

name = { 1: "ABHIJEET", 0: "HERO"}

for face in face_detected:

(x,y,w,h)= face

roi_gray= gray_img[x:x+w,y:y+h]

label,confidencelevel = face_recog.predict(roi_gray)

print(" confidence ", confidencelevel)
print("label " , label)

draw_rect(test_img, face)

predicted_name = name[label]

put_text(test_img,predicted_name, x,y)


resized_img = cv2.resize(test_img, (500,500))

cv2.imshow("ABHIJEET S NICE", resized_img)

cv2.waitKey(0)

cv2.destroyAllWindows()                

Gives the Following Error:
Preparing data...
Data prepared
('Total faces: ', 23)
('Total labels: ', 23)
Predicting images...
Traceback (most recent call last):
File "OpenCV-Face-Recognition-Python.py", line 328, in
predicted_img1 = predict(test_img1)
File "OpenCV-Face-Recognition-Python.py", line 306, in predict
label, confidence = face_recognizer.predict(face)
TypeError: 'int' object is not iterable

You can view my forked code because I made some Changes.

by just commenting LBF and uncommenting FisherFace or EigenFace I am getting the following error

face_recognizer.train(faces, np.array(labels))
cv2.error: OpenCV(3.4.1) /home/majid/opencv_contrib/modules/face/src/eigen_faces.cpp:72: error: (-210) In the Eigenfaces method all input samples (training images) must be of equal size! Expected 143641 pixels, but was 6724 pixels. in function train

I made a virtual environment and installed OpenCV but I found out that I need to install opencv-contrib-python and I'm getting this error message:
Could not install packages due to an EnvironmentError: [WinError 5] Access is denied: 'c:\users\asmah\envs\opencv\Lib\site-packages\cv2\cv2.cp36-win_amd64.pyd'
Consider using the --user option or check the permissions.
I tried using '--user' but I got this error message:
Can not perform a '--user' install. User site-packages are not visible in this virtualenv.
What should I do?

hello！
when I use
ace_recognizer = cv2.face.FisherFaceRecognizer_create()
or
face_recognizer = cv2.face.EigenFaceRecognizer_create()
There is an error：cv2.error: OpenCV(3.4.2) /opt/concourse/worker/volumes/live/9523d527-1b9e-48e0-7ed0-a36adde286f0/volume/opencv-suite_1535558719691/work/opencv_contrib-3.4.2/modules/face/src/fisher_faces.cpp:81: error: (-210:Unsupported format or combination of formats) In the Fisherfaces method all input samples (training images) must be of equal size! Expected 8836 pixels, but was 6724 pixels. in function 'train'

why ?
thank you !

face_recognizer = cv2.face.LBPHFaceRecognizer_create()

AttributeError: module 'cv2.cv2' has no attribute 'face'

the code only detects faces but doesn't recognise the face.if in the test data an unkonwn image is feeded then also it gives the same name

Dear sir.
When I practice face detection by using the code as follows and an error happened as mentioned in Title.
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
face_cascade = cv2.CascadeClassifier('opencv-files/lbpcascade_frontalface.xml')
faces = face_cascade.detectMultiScale(gray, scaleFactor=1.0, minNeighbors=5)
I am sure that the the file "lbpcascade_frontalface.xml" has in the folder "opencv-files" which I created.
by the way, I got "lbpcascade_frontalface.xml" from my colleague because I don't know where to download this file on web.
I am looking forward for your answer, thank you very much.

How can we obtain the corresponding LBP image and histogram from the input image?

OpenCV(4.5.1) C:\Users\appveyor\AppData\Local\Temp\1\pip-req-build-i1s8y2i1\opencv\modules\core\src\lda.cpp:1087: error: (-5:Bad argument) At least two classes are needed to perform a LDA. Reason: Only one class was given! in function 'cv::LDA::lda'

When I add new pictures to the list of faces, it does not recognize the face at all.

Hi,

So, i have one question. In your datasets you used 2 classes... thus, is it possible to just overfit to a single one? We've been testing for a while and i'm not being able to recognize only the person that i want.

Thanks in advance.

In the Fisherfaces method all input samples must be of equal size! Expected 12544 pixels, but was 48400 pixels. In function 'cv::face::Fisherfaces::train'

Hi,

I have one question about the training. Shall I use one image or serveral images to train for one people ?

Hi Rameez, I am having a problem in face recognition.

I used LBPHFaceRecognizer() in training and pre trained HaarCascadeClassifier in face recognition. I used two samples faces and 30 images of each face.

Whenever I tested the model using a new face (that the model has not ben trained for) instead of showing un known it every time shows the name of the first face which I used for the training.

Can you help me in this regard?

How to get similarity scores (by setting threshold??), can I use confidence as similarity score??

sorry but im having trouble understanding if all the code is in the same file. also what do I use for the img argument?

hi, thankyou, its really easy to understand your explaination about the theory of LBPH, but how about the classification, what theory that openCV use, can you explain that to? or do you have any reference that explaining the classification? Thankyou very much

Preparing data...
Data prepared
Total faces: 23
Total labels: 23
Predicting images...
Traceback (most recent call last):
File "/Users/avadhesh/Downloads/opencv-face-recognition-python-master/OpenCV-Face-Recognition-Python.py", line 329, in
predicted_img1 = predict(test_img1)
File "/Users/avadhesh/Downloads/opencv-face-recognition-python-master/OpenCV-Face-Recognition-Python.py", line 309, in predict
label_text = subjects[label]
TypeError: list indices must be integers or slices, not tuple

Process finished with exit code 1