Using the Vision API#

Authentication and Configuration#

  • For an overview of authentication in google-cloud-python, see Authentication.

  • In addition to any authentication configuration, you should also set the GOOGLE_CLOUD_PROJECT environment variable for the project you’d like to interact with. If the GOOGLE_CLOUD_PROJECT environment variable is not present, the project ID from JSON file credentials is used.

    If you are using Google App Engine or Google Compute Engine this will be detected automatically.

  • After configuring your environment, create a Client

>>> from google.cloud import vision
>>> client = vision.Client()

or pass in credentials and project explicitly

>>> from google.cloud import vision
>>> client = vision.Client(project='my-project', credentials=creds)

Annotating an Image#

Annotate a single image#

>>> import io
>>> from google.cloud import vision
>>> client = vision.Client()
>>> with io.open('./image.png', 'rb') as image_file:
...     image = client.image(content=image_file.read())
>>> faces = image.detect_faces(limit=10)
>>> faces[0].landmarks.left_eye.position.x_coordinate
... 1004.8003

Annotate multiple images#

>>> import io
>>> from google.cloud import vision
>>> client = vision.Client()
>>> with io.open('./image.png', 'rb') as image_file:
...     image_one = client.image(content=image_file.read())
>>> image_two = client.image(source_uri='gs://my-storage-bucket/image.jpg')
>>> with client.batch():
...     labels = image_one.detect_labels()
...     faces = image_two.detect_faces(limit=10)

No results returned#

Failing annotations return no results for the feature type requested.

>>> from google.cloud import vision
>>> client = vision.Client()
>>> image = client.image(source_uri='gs://my-storage-bucket/image.jpg')
>>> logos = image.detect_logos(limit=10)
>>> logos
[]

Manual Detection#

You can call the detection method manually.

>>> from google.cloud import vision
>>> from google.cloud.vision.image import Feature
>>> from google.cloud.vision.image import FeatureTypes
>>> client = vision.Client()
>>> image = client.image(source_uri='gs://my-test-bucket/image.jpg')
>>> features = [Feature(FeatureTypes.FACE_DETECTION, 5),
...             Feature(FeatureTypes.LOGO_DETECTION, 3)]
>>> annotations = image.detect(features)

Face Detection#

Detecting a face or faces in an image. For a list of the possible facial landmarks see: https://cloud.google.com/vision/reference/rest/v1/images/annotate#type_1

>>> from google.cloud import vision
>>> client = vision.Client()
>>> image = client.image(source_uri='gs://my-test-bucket/image.jpg')
>>> faces = image.detect_faces(limit=10)
>>> faces[0].landmarks.left_eye.landmark_type
'LEFT_EYE'
>>> faces[0].landmarks.left_eye.position.x_coordinate
1301.2404
>>> faces[0].detection_confidence
0.9863683
>>> faces[0].joy_likelihood
0.54453093
>>> faces[0].anger_likelihood
0.02545464

Label Detection#

Image labels are a way to help categorize the contents of an image. If you have an image with a car, person and a dog it, label detection will attempt to identify those objects.

>>> from google.cloud import vision
>>> client = vision.Client()
>>> image = client.image(source_uri='gs://my-storage-bucket/image.jpg')
>>> labels = image.detect_labels(limit=3)
>>> labels[0].description
'automobile'
>>> labels[0].score
0.9863683

Landmark Detection#

The API will attemtp to detect landmarks such as Mount Rushmore and the Sydney Opera House. The API will also provide their known geographical locations if available.

>>> from google.cloud import vision
>>> client = vision.Client()
>>> image = client.image('./image.jpg')
>>> landmarks = image.detect_landmarks()
>>> landmarks[0].description
'Sydney Opera House'
>>> landmarks[0].locations[0].latitude
-33.857123
>>> landmarks[0].locations[0].longitude
151.213921
>>> landmarks[0].bounding_poly.vertices[0].x_coordinate
78
>>> landmarks[0].bounding_poly.vertices[0].y_coordinate
162

Logo Detection#

Google Vision can also attempt to detect company and brand logos in images.

>>> from google.cloud import vision
>>> client = vision.Client()
>>> image = client.image('./image.jpg')
>>> logos = image.detect_logos(limit=1)
>>> results.logos[0].description
'Google'
>>> logos[0].score
0.9795432
>>> logos[0].bounding_poly.vertices[0].x_coordinate
78
>>> logos[0].bounding_poly.vertices[0].y_coordinate
62

Safe Search Detection#

Detecting safe search properties of an image.

>>> from google.cloud import vision
>>> client = vision.Client()
>>> image = client.image('./image.jpg')
>>> safe_search = image.detect_safe_search()
>>> safe_search.adult
'VERY_UNLIKELY'
>>> safe_search.medical
'UNLIKELY'

Text Detection#

Detecting text with ORC from an image.

>>> from google.cloud import vision
>>> client = vision.Client()
>>> image = client.image('./image.jpg')
>>> texts = image.detect_text()
>>> texts[0].locale
'en'
>>> texts[0].description
'some text in the image'
>>> texts[1].description
'some other text in the image'

Image Properties#

Detecting image color properties.

>>> from google.cloud import vision
>>> client = vision.Client()
>>> image = client.image('./image.jpg')
>>> colors = image.detect_properties()
>>> colors[0].red
244
>>> colors[0].blue
134
>>> colors[0].score
0.65519291
>>> colors[0].pixel_fraction
0.758658