基于注意力机制的音频事件检测

udio event detection has excellent application prospects in fields such as education and security. With the rapid development of deep learning, most of the current research in the field of audio event detection is also based on neural network model. However, few researchers focus on the use of attention mechanism in audio event detection. In this paper, attention mechanism is introduced into the field of audio event detection. In task 5 of dcase2020, this paper uses channel attention and spatial attention to carry out experiments respectively, which proves that the attention mechanism can also play a good effect in the task of audio event detection. Finally, this paper combines channel attention and spatial attention, and improves auprc score and F1 score by 0.66% and 0.75 respectively.