Wake Word Detection - TinyML Project

Introduction

I followed the guide from "Chapter 7: Wake-Word Detection: Building an Application" step by step and successfully build my own application. I selected 2 words from the Speech Commands dataset: "wow" and "happy" as my model's training words. My model is used to recognize the two words.

Approach

• Change the selected words in the code.
• Follow the steps to train the recognition model.
• Freeze the graph.
• Configure the Arduino IDE and TensorFlow Lite Ardiono.
• Deploy the model on Arduino device.
• Test the result.

Result

The model is able to identify the voice input "wow" and "happy".

When the input is "wow", the green LED light will light up.

When the input is "happy", the red LED light will light up.