[Python] Use the Face API of Microsoft Cognitive Services

Microsoft Cognitive Services - Face API https://www.microsoft.com/cognitive-services/en-us/face-api You can detect the position of the face contained in the image, the gender of the person, the age, etc.

Subscription Key Get a Face Preview (free) Subscription Key from Cognitive Services. There are Key1 and Key2, but only Key1 is required.

Hit the API from Python

Version is 2.7. I think it works even with 3. Install requests with pip in advance.

$ pip install requests

Image file version

detect.py

import sys
import requests


url = 'https://api.projectoxford.ai/face/v1.0/detect'
headers = {
    'Content-Type': 'application/octet-stream',
    'Ocp-Apim-Subscription-Key': '[your subscription key]',
}
params = {
    'returnFaceId': 'true',  # The default value is true.
    'returnFaceLandmarks': 'false', # The default value is false.
    'returnFaceAttributes': 'age,gender', # age, gender, headPose, smile, facialHair, and glasses.
}
if __name__ == '__main__':
    argv = sys.argv
    if len(argv) == 1:
        print 'Usage: # python %s [filename]'  % argv[0]
        quit()
    r = requests.post(url ,headers = headers,params = params,data = open(argv[1],'rb'))

    print(r.text)

If you want to detect the face of image.png, execute as follows.

$ python detect.py image.png

Image URL version

When specifying the image by URL, do as follows.

detect.py

import sys
import json
import requests


url = 'https://api.projectoxford.ai/face/v1.0/detect'
image_url = 'http://example.com/image.png'
headers = {
    'Content-Type': 'application/json',
    'Ocp-Apim-Subscription-Key': '[your subscription key]',
}
params = {
    'returnFaceId': 'true',  # The default value is true.
    'returnFaceLandmarks': 'false', # The default value is false.
    'returnFaceAttributes': 'age, gender', # age, gender, headPose, smile, facialHair, and glasses.
}

payload = {
    'url': image_url,
}

if __name__ == '__main__':
    r = requests.post(url ,headers = headers, params = params, data = json.dumps(payload))

    print(r.text)

If you want to detect the face of image_url, execute as follows.

$ python detect.py

Output result

[
    {
        "faceId": "xxxxxxxxxxxxxxxxxxxxxxxx",
        "faceRectangle": {
            "top": 119,
            "left": 177,
            "width": 144,
            "height": 144
        },
        "faceAttributes": {
            "gender": "female",
            "age": 17.9
        }
    }
]

You can see that the result is returned in Json. By changing params, you can also get the position information of the facial part.