Prediction overview

MclassPredict uses a trained classifier context to make class predictions (interferences) on a target. For a trained CNN, segmentation, object detection, or anomaly detection classifier context, your target is either an image or a dataset of images. For a trained tree ensemble classifier context, your target is either a feature (list of values) or a dataset of features.

Specifying a dataset as the target can help training and also can help label your data. For more information, see Analysis, adjustment, and additional settings and Assisted labeling.

The Classification module also lets you import a trained ONNX machine learning model into an ONNX classifier context and use it for prediction. For more information, see ONNX.

Predict engine

The predict engine refers to the hardware (CPU/GPU) with which the prediction is performed. MclassPredict uses the default predict engine established by the Aurora Imaging Configurator utility. You can modify this by calling MclassControl with M_PREDICT_ENGINE or in Aurora Imaging Configurator. Aurora Imaging Library provides an example, PredictEngineSelection.cpp, to help you pick the fastest prediction engine available to you.

[Image: MclassPredictEngineSelectionExample.png]

To run/view this and other examples, use Aurora Imaging Example Launcher.

Note: To take advantage of all available resources, such as GPU prediction using OpenVINO and CUDA, you should install the required Aurora Imaging Library add-on. For more information, see Aurora Imaging Library add-ons and updates.