The IMAGE_EMBEDDINGS action generates vector embeddings for image content stored in Content Lake.
An image embedding is a numerical vector representation of the semantic meaning of an image. These vectors can be used by AI systems to identify visually similar content, support image retrieval, enable recommendation scenarios, and power other embedding-based search and analytics solutions.
The action submits a Content Lake image reference to the Knowledge Enrichment service, generates an embedding vector from the image content, and returns the resulting vector to the workflow.
The input parameters of the action are:
| Parameter | Type | Required | Description |
|---|---|---|---|
| content | Content | Required | The Content Lake document to process. The connector submits the referenced image for
embedding generation, for example:
{
"uri": "hxpr:/1234567890"
}
|
| pollingInitialDelayMs | Integer | Optional | Overrides the initial polling delay in milliseconds. The default value is 1000 milliseconds (1 second). |
| pollingTimeoutSeconds | Integer | Optional | Overrides the timeout period, in seconds, while waiting for processing to complete. |
The output parameters of the action are:
| Parameter | Type | Required | Description |
|---|---|---|---|
| imageEmbeddings | Array | Optional | The generated image embedding represented as an array of numeric vector values, for example: [0.012, -0.034, 0.056] |