Need guidance on automating my trained ML model

Question/Issue: I have ECG data in the form of CSV files, I trained a DNN model in Edge Impulse with this data, before training I performed few preprocessing steps to this data using python. I am happy with the performance of the trained model in edge impulse.

Now the question is, in real time implementation I want to make my ESP32 board to collect data from an ECG sensor and perform preprocessing steps maybe with Micro Python and give them as input to the trained ML model within ESP32. Inorder to achieve this would you recommend:

  1. to build C++ model of the project from edge impulse and work on it to take input from a sensor and to perform pre processing steps before giving them as input to classifier?


  1. would you recommend to simply flash the ML model into ESP32 and to also deploy a micro python script to collect the data from senor and perform preprocessing steps then give them as input to the ML model which is already flashed in the ESP32

In either of the case can you explain in detail how can I achieve this pipeline of communication from a sensor data to perform pre processing and to give it to ML model in ESP32?

Project ID: 367208

Context/Use case:

Hi @yogeshwar_t

Route 1 sounds like the simpler path

Have you got a start on the implementation to share?