Skip to content
Related Articles

Related Articles

Improve Article

Get emotions of images using Microsoft emotion API in Python

  • Difficulty Level : Expert
  • Last Updated : 16 Jul, 2020
Geek Week

The emotions of images like happy, sad, neutral, surprise, etc. can be extracted using Microsoft emotion API for any development purpose.

It is very simple to use and can be called via API through terminal or any of languages like Python or PHP. Microsoft provides free subscription of 30 days for making total of 30,000 requests.
The details of the end points and parameters can be found in the documentation.

# Python script to analyze
# emotion of image
import http.client, urllib.request
import urllib.parse, urllib.error
import base64, sys
import simplejson as json
# replace with subscription_key
# you obtained after registration
subscription_key = '12f29133caf4406493e81b6a31c47c1a'
headers = {
    # Request headers. Replace
    # the placeholder key
    # below with your
    # subscription key.
    'Content-Type': 'application/json',
    'Ocp-Apim-Subscription-Key': subscription_key,
params = urllib.parse.urlencode({
# Replace the URL
# below with the
# URL of the image
# you want to analyze.
body = { 'url': url1 }
newbody =str(body)
    # NOTE: You must use the same region in your REST call as you used to obtain your subscription keys.
    # For example, if you obtained your subscription keys from westcentralus, replace "westus" in the
    # URL below with "westcentralus".
    conn = http.client.HTTPSConnection('')
    conn.request("POST", "/emotion/v1.0/recognize?%s" % params, newbody, headers)
    response = conn.getresponse()
    data =
    parsed = json.loads(data)
    print ("Response:")
    print (json.dumps(parsed, sort_keys=True, indent=2))
    # the emotion of image
    # will the max value of
    # any emotion obtained
    # from the different
    # scores of each emotion
    val = parsed[0]["scores"]
    res = max(val, key = val.get)
    print ("\nEmotion :: ",res)
except Exception as e:

The sample project using this api is available on SnapLook

GeeksforGeeks LIVE courses

My Personal Notes arrow_drop_up
Recommended Articles
Page :