Acoustic identification of individual animals can be improved by leveraging hierarchical contrastive learning to create robust representations that preserve the hierarchical relationships between species and taxa.
AudioProtoPNet은 새소리 분류를 위한 해석 가능한 딥러닝 모델로, 새소리 데이터의 특징적인 부분을 프로토타입으로 학습하여 분류 결과에 대한 설명을 제공한다.
AudioProtoPNet is an interpretable deep learning model that can accurately classify bird species from audio recordings by learning and identifying prototypical sound patterns.
Deep learning models in avian bioacoustics face challenges due to inconsistent research practices. BirdSet benchmark aims to address these issues by providing a unified framework for classifying bird vocalizations.
Creating a unified benchmark for classifying bird vocalizations in avian bioacoustics.
Mix2 framework effectively addresses multi-label imbalanced classification challenges in bioacoustics.
The author explores the application of Wavelet Scattering Transform (WST) in bioacoustics, outperforming existing methods by 6% using WST and 8% using Mel spectrogram preprocessing, achieving a top accuracy of 96%.