Dynamic noise filtering for multi-class classification of beehive audio data
作者：
Highlights：
• A multi-class bee audio classification task with poor state-of-the-art performance.
• A noise filtering method improving the performance of bee audio data classification.
• Large-scale benchmark data for beehive audio data analytics research.
• Comparison of various audio feature extraction and noise filtering approaches.
• Highly accurate multi-label bee audio classification with very simple models.
摘要
•A multi-class bee audio classification task with poor state-of-the-art performance.•A noise filtering method improving the performance of bee audio data classification.•Large-scale benchmark data for beehive audio data analytics research.•Comparison of various audio feature extraction and noise filtering approaches.•Highly accurate multi-label bee audio classification with very simple models.
论文关键词：Audio data analysis,Spectrogram,Noise filtering,Audio feature extraction,Apiculture
论文评审过程：Received 13 July 2022, Revised 12 September 2022, Accepted 14 September 2022, Available online 20 September 2022, Version of Record 1 October 2022.
论文官网地址：https://doi.org/10.1016/j.eswa.2022.118850
