Does Deep Learning HDL support 1-D classification methods in MATLAB R2022a?

All the details about supported example models for the Deep Learning HDL Toolbox seem to suggest that the only input type supported by the Deep Learning HDL toolbox is image data. I need the input to be an array of numbers of various sizes/types. Does Deep Learning HDL provide support for these inputs as well?

Answer 1

[MathWorks Support Team]

Deep Learning HDL Toolbox actually has an example workflow with non-image inputs to perform data classification. The example looks at frames of length 1024, and determines the type of modulation that was applied to the signal after sources of interference are added in. Please refer to the following documentation page for more information:

 

https://www.mathworks.com/help/releases/R2022a/deep-learning-hdl/ug/modulation-classification-by-using-FPGA.html

In this example, although the use of an Image Input Layer, the input is of size 1x1024. So despite it looks like it performs a 2-D image processing, it actually performs a 1-D classification.

 

In MATLAB R2022b, the Deep Learning HDL Toolbox supports sequential input networks like LSTM. Other input layers such as the Feature Input Layer and Sequential Input Layer are also included in MATLAB R2022b, as observed in the listings of supported layers  below:

https://www.mathworks.com/help/releases/R2022b/deep-learning-hdl/ug/supported-networks-layers-boards-and-tools.html