ni.com is currently undergoing scheduled maintenance.

Some services may be unavailable at this time. Please contact us for help or try again later.

Example Code

ASK A QUESTION

by: Last Edited by:

Products and Environment

To download NI software, including the products shown below, visit ni.com/downloads.

This example demonstrates how to use an ANN (artificial neural network) in LabVIEW.

It includes a wrapper over the ONNX Runtime, allowing you to:

Image processing is implemented using the following logic:

Stage 1: Detection (YOLOv8n ANN):

Convert the [1024, 1024] image to a [1, 3, 1024, 1024] tensor and feed it to the ANN input
Convert the data received from the ANN to a readable form, filter by threshold and NMS
Display the detection result

Stage 2: Rotation angle compensation (custom ANN):

Image segmentation (based on the results of stage 1)
Image resizing [x, x] -> [512, 512]
Convert the [512, 512] image to a [1, 3, 512, 512] tensor and feed it to the ANN input
Convert the [sin, cos] data received from the ANN to an angle and compensate for it
Display the result (segmentation, scaling, rotation)

Stage 3: Identification (ANN EfficientNet-B4):

Image resizing [512, 512] -> [380, 380]
Transform image [380, 380] into tensor [1, 3, 380, 380] and feed it to ANN input
Transform data received from ANN into probabilities for each of the two classes [Avers, Revers]
Display result

Example using VIs