Identification of Emergency Vehicle Sirens Based on Deep Learning
Shu‐Min Tsai, Ming‐Lin Chuang, Wen-Pu Yang · 2024
Ambulances, police cars and fire trucks are emergency vehicles that need to go quickly to deal with things. The road should be as clear as possible to allow them to pass quickly. This research uses deep learning to identify emergency vehicles and provide warning applications. The deep learning architecture uses the CNN module and uses MFCC as the input pattern. When the signal dimension is [20, 22] and the sampling rate is 44100, the best results are obtained, with an accuracy rate as high as 98%. The precision, recall and F1 score of test file are about equal 1.