Analog Signal Processing for creating a Deep Learning classifier model

AI Discussions
1

I am starting to work on a project where it is required to build a deep learning classifier model for gesture classification. The input data consists of analog voltage signal in csv file format. How should I start with the project, as in what pre-processing should I do, feature extraction and model pipeline.

tek0001
tek0001800×480 6.22 KB

Sample image of the signal (voltage) from oscilloscope.

2

Sequences of voltage samples are typically an application for a Recurrent Neural Network (such as an LSTM or GRU).

You’ll need a way to attach labels to the data so that the gesture is identified.

I recommend you study the Deep Learning Specialization, Course 5, which is all about RNN’s.