Environmental Sound Recognition based on a STM32 SensorTile board

Recently, with the development of the artificial intelligence (AI). The application of Deep Learning is expending into different field. Image recognition, personalized recommendations in search engines, Natural Language Processing (NLP) and Speech Recognition are based on the Deep Learning. The continued development and integration of these applications can also lead to a more convenient lifestyle. Including Autonomous Vehicle, Facial Recognition and more. Based on the structure of Deep Learning. We transfer the data of the ambient sound into the Log-Mel spectrogram. Then using the Log-Mel spectrogram to train the Convolutional Neural Network (CNN) model. After this, we deployed the trained model onto the STM32 SensorTile board for experimentation.