How do I fix “RuntimeError: Unable to open shape_predictor_68_face_landmarks.dat” while using google colab?

问题

I am writing the following Python code in Google Colaboratory and get an error:

Code:

import cv2
import dlib

cap = cv2.VideoCapture(0)

hog_face_detector = dlib.get_frontal_face_detector()

dlib_facelandmark = dlib.shape_predictor("shape_predictor_68_face_landmarks.dat")

while True: _,
frame = cap.read()
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

faces = hog_face_detector(gray)
for face in faces:

    face_landmarks = dlib_facelandmark(gray, face)

    for n in range(0, 16):
        x = face_landmarks.part(n).x
        y = face_landmarks.part(n).y
        cv2.circle(frame, (x, y), 1, (0, 255, 255), 1)


cv2.imshow("Face Landmarks", frame)

key = cv2.waitKey(1)
if key == 27:
    break
cap.release()
cv2.destroyAllWindows()

Error:

RuntimeError Traceback (most recent call last) in () 6 hog_face_detector = dlib.get_frontal_face_detector() 7 ----> 8 dlib_facelandmark = dlib.shape_predictor("shape_predictor_68_face_landmarks.dat") 9 10 while True:

RuntimeError: Unable to open shape_predictor_68_face_landmarks.dat

回答1:

Your notebook is unable to open the file shape_predictor_68_face_landmarks.dat. This may happen because the file isn't uploaded on your notebook, or because the path you specified in dlib.shape_predictor("shape_predictor_68_face_landmarks.dat") is wrong.

The following is an edit of your code that automatically downloads the bzip2 file, extracts it and sets it as your shape predictor. You can use different .dat files by changing the link of !wget.

Cell 1:

!wget   http://dlib.net/files/shape_predictor_68_face_landmarks.dat.bz2 # DOWNLOAD LINK

!bunzip2 /content/shape_predictor_68_face_landmarks.dat.bz2

datFile =  "/content/shape_predictor_68_face_landmarks.dat"

Cell 2:

import cv2 
import dlib

cap = cv2.VideoCapture(0)

hog_face_detector = dlib.get_frontal_face_detector()

dlib_facelandmark = dlib.shape_predictor(datFile)

while True: _, 
frame = cap.read() 
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

faces = hog_face_detector(gray)
for face in faces:

    face_landmarks = dlib_facelandmark(gray, face)

    for n in range(0, 16):
        x = face_landmarks.part(n).x
        y = face_landmarks.part(n).y
        cv2.circle(frame, (x, y), 1, (0, 255, 255), 1)


cv2.imshow("Face Landmarks", frame)

key = cv2.waitKey(1)
if key == 27:
    break
cap.release() 
cv2.destroyAllWindows()

Moreover, be aware that if you use cv2.VideoCapture(0) on a Jupyter notebook to open a camera it won't work, since the code is running on some remote server, and not on you computer. Look at the snippet here for an example on how to access your local webcam in Colab.

来源：https://stackoverflow.com/questions/64643440/how-do-i-fix-runtimeerror-unable-to-open-shape-predictor-68-face-landmarks-dat

标签

python

OpenCV

image-processing

webcam

facial-identification