| Info |
|---|
| Note |
This plugin is no longer available as of July 26, 2024. |
This transform transformation leverages the CognitiveJ library, and in turn the Azure Cognitive APIs to extract faces and other metadata, including emotions, from a set of images. It is usually used in conjunction with the Whole File Reader plugin since it requires the entire contents of each image to be loaded into a single message and passed into the transform. Due to this, there may be memory issues when loading large images. This leverages the Face API and the Emotion API specifically for this task.
Use Case
A developer is analyzing a large number of files and would like to identify faces in those photos to see how many men and women are in the photos. The user can combine this plugin with the whole file reader and filter using wrangler plugins based on the data extracted.
Configuration
Property | Macro Enabled? | Description |
|---|---|---|
Source Field | No | Required. This is the name of the field on the input record containing the image file. It must be of type bytes and it must contain the entire contents of the image file. |
Azure Face API Key | Yes | Required. This key is obtained from the Azure Portal after enabling the Face API. |
Azure Emotion API Key | Yes | Required. This key is obtained from the Azure Portal after enabling the Emotion API. |
Continue Processing If There Are Errors? | Yes | Required. Indicates if the pipeline should continue if processing a single image fails. Default is false. |
Usage Notes
This plugin requires an Azure account as well as having the Face and Emotion APIs enabled on the account. This plugin will make two calls per image, and pricing is based on requests to the service. Please see the Pricing details for more information about pricing.
...