RGB float image for deep learning

[tomsze]

I would like to classify two objects using mobilenet (input shape 224x224x3). I have trained the model in Keras and converted to .pb format. To make the training process easier, I preprocessed the input RGB image pixel value to [-1..1] range. However, I was not able to create a RGB float image in labview to input to the network. Is there a way to create this kind of image for Labview's DL Model Set Input.vi?

By the way, what is the proper way of creating a array for Labview's DL Model Set Input.vi? I tried, but the result was not as expected.

[tomsze · Accepted solution]

I have found the solution of usng array as input.

Labview 4D array is slightly different from python numpy 4D array so it needs to be rearranged.

Labview array for DL