Microsoft just released a Computer Vision API. I tried it with a couple of Unidentified objects of Wikipedia and was not really impressed.
Here is how to use it:
#!/usr/bin/env python
"""Example how to use the Microsoft Vision API."""
import httplib
import urllib
import json
headers = {"Content-type": "application/json"}
params = urllib.urlencode(
{
# Specify your subscription key
"subscription-key": "yourhexadecimalone11111111111111",
# Specify values for optional parameters, as needed
"visualFeatures": "All",
}
)
try:
image_url = "http://www.martin-thoma.de/bilder/Martin_Thoma_web_thumb.jpg"
conn = httplib.HTTPSConnection("api.projectoxford.ai")
conn.request(
"POST", "/vision/v1/analyses?%s" % params, "{'Url': '%s'}" % image_url, headers
)
response = conn.getresponse()
data = response.read()
data = json.loads(data)
print(json.dumps(data, sort_keys=True, indent=2))
conn.close()
except Exception as e:
print("[Errno {0}] {1}".format(e.errno, e.strerror))
which gives
{
"adult": {
"adultScore": 0.01073759701102972,
"isAdultContent": false,
"isRacyContent": false,
"racyScore": 0.015348214656114578
},
"categories": [
{
"name": "people_",
"score": 0.8359375
}
],
"color": {
"accentColor": "4C5F37",
"dominantColorBackground": "Green",
"dominantColorForeground": "Green",
"dominantColors": [
"Green"
],
"isBWImg": false
},
"faces": [
{
"age": 28,
"faceRectangle": {
"height": 49,
"left": 53,
"top": 42,
"width": 49
},
"gender": "Male"
}
],
"imageType": {
"clipArtType": 0,
"lineDrawingType": 0
},
"metadata": {
"format": "Jpeg",
"height": 200,
"width": 134
},
"requestId": "7f0af611-1750-4fa2-aa47-a03db286f6a7"
}
See also
- Cognitive Services: You have to apply there. 5,000 transactions per month, 20 per minute is for free.
- The 86-category concept
- Computer Vision API - v1.0