Description
This component allows you to specify an inference model to be performed on an image file.
| Output payload | Description |
|---|---|
cv.Payload | The result of the inference run is output |
Component Property
| property name | Description |
|---|---|
| image file | Specifies the file path of a picture saved as a file |
| Inference model | Specify the model name of the inference to be performed |
| Filter1 | Select to filter inference results |
| filter1 | select if you want to filter inference results |
| filter2 | select if you want to filter inference results |
| Condition | Specify the condition for filter 2 |
| Conversion | Specify a conversion |
| Draw | Specify if you want to save the inference result as an image |
| Copy Destination | Specify where to save the image if "Copy" is specified in "Draw" |
Filter
This is used when you want to extract only the results that match the condition from the inference results.
| Filter | Description |
|---|---|
| None | Do not use filter |
| Default of the model | Use the filter defined in the model |
| highest n predictions | retrieve the n data specified by the condition from the higher of multiple predictions |
| the lowest n predictions | retrieve the n data specified by the condition from the lowest of the multiple predictions |
| predictions scored higher than | retrieve the n data specified by the condition from the lowest of the multiple predictions |
| predictions cored lower than | retrieve data from multiple forecasts whose score is less than x specified by the condition |
| prediction labeled with | retrieve data from multiple predictions where the label contains a string specified by the condition |
Conversion
Use this to convert the result of inference.
| Conversion | Description |
|---|---|
| None | Do not use transformation |
| Default of the model | Use the transformation defined in the model |
| number | get the number of inference results |
| first label | get the first label of the inference result |
| all labels | retrieve all labels in the inference result |
Draw
Use to save the inference result as an image.
| Draw | Description |
|---|---|
| None | Do not save the result of inference as an image |
| Original | Save the result of inference as an original image |
| Copy | Create another file in the destination of the copy and save the result of the inference |
Use in combination with CameraTakesPhotos component
You can use the CameraTakesPhotos component when executing an action to perform inference on a photo saved as a file.
Place a Split component between the CameraTakesPhotos component and the ExecuteInferenceModel component as shown in this example.
Since the output of the CameraTakesPhotos component is an array of filenames of the saved photos, the Split component splits the array into individual items to be processed each in the subsequent ExecuteInferenceModel component.