This project presents the implementation of a keyword spotter using the Nano 33 BLE Arduino Board. The aim is to develop an efficient wake word detection system, essential for many speech recognition applications. Wake word detection involves recognizing specific phrases or words, in this case, "yes" and "computer," to trigger the system into action.
Dataset: Collection of audio samples from myself and john smith. Target Wake Words: "Yes" and "Computer". Model Training: A deep learning model trained on the collected dataset to recognize the wake words.
Board Used: Nano 33 BLE Arduino Board. Real-Time Monitoring: Continuous monitoring of audio input for wake word detection.
Final accuracy in software of 92% on training and 87% on validation
