Augmented AI Crowd HTML Elements

A widget to enable human review of a Amazon Textract document analysis result.

Attributes

The following attributes are supported by this element.

header

This is the text that is displayed as the header.

src

This is a link to the image to be analyzed by the worker.

initialValue

This sets initial values for attributes found in the worker UI.

The following is an example of an initialValue input:

[
            {
                "blockType": "KEY_VALUE_SET",
                "confidence": 38.43309020996094,
                "geometry": {
                    "boundingBox": {
                        "width": 0.32613086700439453,
                        "weight": 0.0942094624042511,
                        "left": 0.4833833575248718,
                        "top": 0.5227988958358765
                    },
                    "polygon": [
                        {"x": 0.123, "y": 0.345}, ...
                    ]
                }
                "id": "8c97b240-0969-4678-834a-646c95da9cf4",
                "relationships": [
                    {
                        "type": "CHILD",
                        "ids": [
                            "7ee7b7da-ee1b-428d-a567-55a3e3affa56",
                            "4d6da730-ba43-467c-a9a5-c6137ba0c472"
                        ]
                    },
                    {
                        "type": "VALUE",
                        "ids": [
                            "6ee7b7da-ee1b-428d-a567-55a3e3affa54"
                        ]
                    }
                ],
                "entityTypes": [
                    "KEY"
                ],
                "text": "Foo bar"
            },
]

blockTypes

This determines the kind of analysis the workers can do. Only KEY_VALUE_SET is currently supported.

keys

This specifies new keys and the associated text value the worker can add. The input values for keys can include the following elements:

The following is an example of an input for keys.

[
      {
        importantFormKey: 'Address',
        importantFormKeyAliases: [
          'address',
          'Addr.',
          'Add.',
        ]
      },
      {
        importantFormKey: 'Last name',
        importantFormKeyAliases: ['Surname']
      }
]

no-key-edit

This prevents the workers from editing the keys of annotations passed through initialValue. This prevents workers from editing the keys that have been detected on your documents. This is required.

no-geometry-edit

This prevents workers from editing the polygons of annotations passed through initialValue. For example, this would prevent the worker from editing the bounding box around a given key. This is required.

Element Hierarchy

This element has the following parent and child elements.

Regions

The following regions are supported by this element. You can use custom HTML and CSS code within these regions to format your instructions to workers. For example, use the short-instructions section to provide good and bad examples of how to complete a task.

full-instructions

General instructions about how to work with the widget.

short-instructions

Important task-specific instructions that are displayed in a prominent place.

Example of a Worker Template Using the crowd Element

An example of a worker template using this crowd element would look like the following.

Output

The following is a sample of the output from this element. You can find a detailed explanation of this output in the Amazon Textract AnalyzeDocument API documentation.

{
  "AWS/Textract/AnalyzeDocument/Forms/V1": {
    blocks: [
      {
        "blockType": "KEY_VALUE_SET",
        "id": "8c97b240-0969-4678-834a-646c95da9cf4",
        "relationships": [
          {
            "type": "CHILD",
            "ids": ["7ee7b7da-ee1b-428d-a567-55a3e3affa56", "4d6da730-ba43-467c-a9a5-c6137ba0c472"]
          },
          {
            "type": "VALUE",
            "ids": ["6ee7b7da-ee1b-428d-a567-55a3e3affa54"]
          }
        ],
        "entityTypes": ["KEY"],
        "text": "Foo bar baz"
      }
    ]
  }
}

A widget to enable human review of an Amazon Rekognition image moderation result.

Attributes

The following attributes are supported by this element.

header

This is the text that is displayed as the header.

src

This is a link to the image to be analyzed by the worker.

categories

This supports categories as an array of strings or an array of objects where each object has a name field.

If the categories come in as objects, the following applies:

If the categories come in as strings, the following applies:

exclusion-category

By setting this attribute you create a button underneath the categories in the UI.

Element Hierarchy

This element has the following parent and child elements.

Amazon Regions

The following Amazon Regions are supported by this element. You can use custom HTML and CSS code within these Regions to format your instructions to workers. For example, use the short-instructions section to provide good and bad examples of how to complete a task.

full-instructions

General instructions about how to work with the widget.

short-instructions

Important task-specific instructions that are displayed in a prominent place.

Example Worker Template with the crowd Element

An example of a worker template using the crowd element would look like the following.

Output

The following is a sample of the output from this element. For details about this output, see Amazon Rekognition DetectModerationLabels API documentation.

{
  "AWS/Rekognition/DetectModerationLabels/Image/V3": {
    "ModerationLabels": [
        { name: 'Gore', parentName: 'Violence' },
        { name: 'Corpses', parentName: 'Violence' },
    ]
  }
}