Audio Based Data Collection Techniques for Hand Gesture Recognition
Elec491 - Electrical Engineering Design Project
In this project we combined sound and hand gestures and offered a classification model for hand gesture recognition to be used in a smart home automation project. To optimize cost, complexity and accuracy for acquiring sound from solid surfaces, Before constructing the model we examined different three different microphones based on : MEMS, Electret, and Piezo Disks. We focused on piezo discs because of their superior ability to transform mechanical energy into electrical energy. After data acquisition a pre trained CNN nertwork ( VGG16) was used for model construction.