Inference Action

The Inference Action runs an inference on a selected model with a specific image and receive a JSON response containing the inference data.


Parameters

Parameter Description
Model Name Model Name to be used for inference. The available model names can be selected from a dropdown list.
Inference Location

Determines how the image data is handled based on the location of the deviceWISE Visual Inspection Instance. Options are:

  • Remote: For instances where DWVI is not on the same machine as deviceWISE, the entire image is sent to DWVI for processing.

  • Local: For instances where DWVI and deviceWISE are co-located on the same machine, only a reference to the image is needed, optimizing data handling.
Image Input Source Source from which the image will be provided. Options are: File or Variable.
Delete Image Determines if the image file will be deleted, and if so, when. Options are: Never, Always, On Success. This options are only available if the Image Input Source selected is File.

Input Variables

Parameter Description
Input Image

Specify the image's location on the deviceWISE machine.

Generate Caption (Legacy) A Boolean option to decide whether to overlay annotations directly on the image or not.
Confidence Threshold A float value that allows users to set a minimum confidence level for object detection, ensuring only relevant results are returned.

Output Variables

Parameter Description
Status Indicates the response status of the request made from deviceWISE, providing insight into the success or failure of the operation.
Inference Result Delivered as a JSON string, this data can be integrated into deviceWISE View for further analysis or action triggers.