Detect labels in an image by using the command line

Detect labels in an image by using the command line

This page shows you how to send three feature detection and annotation requests to the Vision API using the REST interface and the curl command.

Vision API enables easy integration of Google vision recognition technologies into developer applications. You can send image data and desired feature types to the Vision API, which then returns a corresponding response based on the image attributes you are interested in. For more information about the feature types offered, see the List of all Vision API features.

Before you begin

1. Sign in to your Google Cloud account. If you're new to Google Cloud, create an account to evaluate how our products perform in real-world scenarios. New customers also get $300 in free credits to run, test, and deploy workloads.
2. Install the Google Cloud CLI.

3. To initialize the gcloud CLI, run the following command:

   gcloud init

4. Create or select a Google Cloud project.

   Note: If you don't plan to keep the resources that you create in this procedure, create a project instead of selecting an existing project. After you finish these steps, you can delete the project, removing all resources associated with the project.

   • Create a Google Cloud project:

     gcloud projects create PROJECT_ID

     Replace PROJECT_ID with a name for the Google Cloud project you are creating.

   • Select the Google Cloud project that you created:

     gcloud config set project PROJECT_ID

     Replace PROJECT_ID with your Google Cloud project name.

5. Make sure that billing is enabled for your Google Cloud project.

6. Enable the Vision API:

   gcloud services enable vision.googleapis.com

7. Grant roles to your user account. Run the following command once for each of the following IAM roles: roles/storage.objectViewer

   gcloud projects add-iam-policy-binding PROJECT_ID --member="user:USER_IDENTIFIER" --role=ROLE

   • Replace PROJECT_ID with your project ID.

   • Replace USER_IDENTIFIER with the identifier for your user account. For example, user:[e-mail].

   • Replace ROLE with each individual role.
8. Install the Google Cloud CLI.

9. To initialize the gcloud CLI, run the following command:

   gcloud init

10. Create or select a Google Cloud project.

    Note: If you don't plan to keep the resources that you create in this procedure, create a project instead of selecting an existing project. After you finish these steps, you can delete the project, removing all resources associated with the project.

    • Create a Google Cloud project:

      gcloud projects create PROJECT_ID

      Replace PROJECT_ID with a name for the Google Cloud project you are creating.

    • Select the Google Cloud project that you created:

      gcloud config set project PROJECT_ID

      Replace PROJECT_ID with your Google Cloud project name.

11. Make sure that billing is enabled for your Google Cloud project.

12. Enable the Vision API:

    gcloud services enable vision.googleapis.com

13. Grant roles to your user account. Run the following command once for each of the following IAM roles: roles/storage.objectViewer

    gcloud projects add-iam-policy-binding PROJECT_ID --member="user:USER_IDENTIFIER" --role=ROLE

    • Replace PROJECT_ID with your project ID.

    • Replace USER_IDENTIFIER with the identifier for your user account. For example, user:[e-mail].

    • Replace ROLE with each individual role.

Make an image annotation request

After completing the Before you begin steps you can use the Vision API to annotate an image file.

In this example you use curl to send a request to the Vision API using the following image:

Cloud Storage URI:

gs://cloud-samples-data/vision/using_curl/shanghai.jpeg

HTTPS URL:

https://console.cloud.google.com/storage/browser/cloud-samples-data/vision/using_curl/shanghai.jpeg

Image credit: Steve Long on Unsplash.

Create the request JSON

The following request.json file demonstrates how to request three images:annotate features and limit the results in the response.

Create the JSON request file with the following text, and save it as a request.json plain text file in your working directory:

request.json

{
  "requests": [
    {
      "image": {
        "source": {
          "imageUri": "gs://cloud-samples-data/vision/using_curl/shanghai.jpeg"
        }
      },
      "features": [
        {
          "type": "LABEL_DETECTION",
          "maxResults": 3
        },
        {
          "type": "OBJECT_LOCALIZATION",
          "maxResults": 1
        },
        {
          "type": "TEXT_DETECTION",
          "maxResults": 1,
          "model": "builtin/latest"
        }
      ]
    }
  ]
}

Field value details

• image.source.gcsImageUri - Indicates the image stored in a Cloud Storage bucket. You change this request to image.source.imageUri for a publicly available URI, or image.content to pass a base64 encoded string representation of an image.
• features - An object representing a specific feature type. You can request multiple feature types for a single image.
• type - The enum value specifying a feature.
• maxResults (optional) - A limiting value on the results returned.
• model (optional) - If applicable you can specify either builtin/stable (the default if unset) or builtin/latest to choose your model. Refer to the Release notes topic for a list of recently updated models.

Send the request

You use curl and the body content from request.json to send the request to the Vision API. Enter the following on your command line:

curl -X POST \
    -H "Authorization: Bearer $(gcloud auth print-access-token)" \
    -H "x-goog-user-project: PROJECT_ID" \
    -H "Content-Type: application/json; charset=utf-8" \
    https://vision.googleapis.com/v1/images:annotate -d @request.json

• The sample curl command uses the gcloud auth print-access-token command to get an authentication token.
• Note that to pass a filename to curl you use the -d option (for "data") and precede the filename with an @ sign. This file should be in the same directory in which you execute the curl command.

Interpret the response

You should see a JSON response similar to the one below.

The request JSON body specified maxResults for each annotation type. Consequently, you will see the following in the response JSON:

• three labelAnnotations results
• one textAnnotations result (shortened for clarity)
• one localizedObjectAnnotations result

Response

Note: Zero coordinate values omitted. When the API detects a coordinate ("x" or "y") value of 0, that coordinate is omitted in the JSON response. Thus, a response with a bounding poly around the entire image would be
[{},{"x": 100},{"x": 100,"y": 100},{"y": 100}] for an image that is 100 pixels by 100 pixels. For more information, see the API Reference documentation.

{
  "responses": [
    {
      "labelAnnotations": [
        {
          "mid": "/m/09g5pq",
          "description": "People",
          "score": 0.9504782,
          "topicality": 0.9504782
        },
        {
          "mid": "/m/01c8br",
          "description": "Street",
          "score": 0.8911568,
          "topicality": 0.8911568
        },
        {
          "mid": "/m/079bkr",
          "description": "Mode of transport",
          "score": 0.89089024,
          "topicality": 0.89089024
        }
      ],
      "textAnnotations": [
        {
          "locale": "zh",
          "description": "牛牛面馆\n",
          "boundingPoly": {
            "vertices": [
              {
                "x": 159,
                "y": 212
              },
              {
                "x": 947,
                "y": 212
              },
              {
                "x": 947,
                "y": 354
              },
              {
                "x": 159,
                "y": 354
              }
            ]
          }
        },
        ...
      ],
      "fullTextAnnotation": {
        "pages": [
          {
            ...
                "paragraphs": [
                  {
                    ...
                    "words": [
                      {
                        ...
                        "symbols": [
                          {
                            ...
                ],
                "blockType": "TEXT"
              }
            ]
          }
        ],
        "text": "牛牛面馆\n"
      },
      "localizedObjectAnnotations": [
        {
          "mid": "/m/01g317",
          "name": "Person",
          "score": 0.94413143,
          "boundingPoly": {
            "normalizedVertices": [
              {
                "x": 0.26063988,
                "y": 0.46869153
              },
              {
                "x": 0.40736017,
                "y": 0.46869153
              },
              {
                "x": 0.40736017,
                "y": 0.8957791
              },
              {
                "x": 0.26063988,
                "y": 0.8957791
              }
            ]
          }
        }
      ]
    }
  ]
}

Label detection results

1. description: "People", score: 0.950
2. description: "Street", score: 0.891
3. description: "Mode of transport", score: 0.890

Image credit: Steve Long on Unsplash (annotations added).

Text detection results

• text: 牛牛面馆\n
• vertices: (x: 159, y: 212), (x: 947, y: 212), (x: 947, y: 354), (x: 159, y: 354 )

Image credit: Steve Long on Unsplash (annotations added).

Object detection results

• name: "Person", score: 0.944
• normalized vertices: (x: 0.260, y: 0.468), (x: 0.407, y: 0.468), (x: 0.407, y: 0.895), (x: 0.260, y: 0.895)

Image credit: Steve Long on Unsplash (annotations added).

Congratulations! You've sent your first request to the Vision API.

Clean up

To avoid incurring charges to your Google Cloud account for the resources used on this page, delete the Google Cloud project with the resources.

Optional: Revoke credentials from the gcloud CLI.

gcloud auth revoke

What's next

• See a list of all feature types and their uses.
• Get started with the Vision API in your language of choice by using a Vision API Client Library.
• Use the How-to guides to learn more about specific features, see example annotations, and get annotations for an individual file or image.
• Learn about batch image and file (PDF/TIFF/GIF) annotation.
• Browse through a comprehensive list of client library code samples.