An unofficial python wrapper for the Facebook face recognition endpoint
- Install fbrecog from pip
pip install fbrecog - Now simply import the recognize method from fbrecog module
from fbrecog import recognize - The recognize method takes 4 input args and returns an array of recognitions.
- To get the access token simply go to https://developers.facebook.com/tools/explorer and get a user access token with user_photos, publish_actions and user_posts permissions.
- Get your Facebook cookie and fb_dtsg token as follows:
-
Go to your Facebook profile.
-
Open chrome dev tools by
Right Click > Inspect -
Upload any picture. As it gets uploaded monitor the Network tab for 'dpr?' endpoint.
-
Click on it. Scroll down to Request Header. Copy the entire cookie string.
-
Scroll further down to Form Data. Copy the value of fb_dtsg parameter.
- Call the recognize method with the following parameters.
from pprint import pprint
from fbrecog import FBRecog
path = '1.jpg' # Insert your image file path here
path2 = '2.jpg' # Insert your image file path here
access_token = '#######' # Insert your access token obtained from Graph API explorer here
cookie = '###' # Insert your cookie string here
fb_dtsg = '###' # Insert the fb_dtsg parameter obtained from Form Data here.
# Instantiate the recog class
recog = FBRecog(access_token, cookies, fb_dtsg)
# Recog class can be used multiple times with different paths
print(recog.recognize(path))
print(recog.recognize(path2))
# Call recognize_raw to get more info about the faces detected, including their positions
pprint(recog.recognize_raw(path), indent=2)