Package org.anchoranalysis.image.inference

Interface ImageInferenceModel<T>

Type Parameters:

T - tensor-type that is both inputted to (representing an image), and outputted from the model during inference.

All Superinterfaces:

AutoCloseable, InferenceModel

All Known Implementing Classes:

OnnxModel, OpenCVModel

public interface ImageInferenceModel<T>
extends InferenceModel

A model used for inference that accepts an image as an input.

A model should always be closed, when it is no longer in use.

Author:

Owen Feehan

Method Summary

Modifier and Type	Method	Description
<S> S	performInference(T input, String inputName, List<String> outputIdentifiers, CheckedFunction<List<T>,S,OperationFailedException> convertOutput)	Performs inference on a single-input, to create an output.

Methods inherited from interface org.anchoranalysis.inference.InferenceModel

close

Method Details

performInference

<S> S performInference(T input,
 String inputName,
 List<String> outputIdentifiers,
 CheckedFunction<List<T>,S,OperationFailedException> convertOutput)
                throws OperationFailedException

Performs inference on a single-input, to create an output.

Type Parameters:

S - the data-type the output is exposed as.

Parameters:

input - the input for inference.

inputName - the name associated with input in the model.

outputIdentifiers -

convertOutput - converts the output to type <S>.

Returns:

the converted output.

Throws:

OperationFailedException - if the inference cannot successfully complete.