Image Classification

Purpose of the Node

The Image Classification node predicts the class of an input image using an ONNX model. It applies default transformations to the input image such as resizing, center cropping, normalizing, and scaling pixel values to ensure compatibility with the ONNX model. The node also supports optional softmax activation to scale the output logits to probabilities.

Pins

Pin Name	Pin Description	Pin Type	Value Type
Start	Initiate Execution	Execution	-
Model	ONNX Model Session	Struct	NodeOnnxSession
Image	Image Object	Struct	NodeImage
Mean	Image Mean for Normalization (per channel)	Float	Array
Std	Image Standard Deviation for Normalization (per channel)	Float	Array
Crop	Center Crop Percentage	Float	-
Softmax?	Scale Outputs with Softmax	Boolean	-
End	Done with the Execution	Execution	-
Predictions	Class Predictions	Struct	Array