DOAJ Open Access 2025

A Multi-Time-Frequency Feature Fusion Approach for Marine Mammal Sound Recognition

Xiangxu Meng Xin Liu Yinan Xu Yujing Wu Hang Li +3 lainnya

Abstrak

Accurate acoustic identification of marine mammals is vital for monitoring ocean health and human impacts. Existing methods often struggle with limited single-feature representations or suboptimal fusion of multiple features. This paper proposes an Evaluation-Adaptive Weighted Multi-Head Fusion Network that integrates CQT and STFT features via a dual-branch ResNet architecture. The model enhances intra-branch features using channel attention and adaptive weighting of each branch based on its validation accuracy during training. Experiments on the Watkins Marine Mammal Sound Database show that the proposed method achieves superior performance, reaching 96.05% accuracy and outperforming baseline and attention-based fusion models. This approach offers an effective solution for multi-feature acoustic recognition in complex underwater environments.

Topik & Kata Kunci

Naval architecture. Shipbuilding. Marine engineering Oceanography

Penulis (8)

Xiangxu Meng

Xin Liu

Yinan Xu

Yujing Wu

Hang Li

Kye-Won Kim

Suya Liu

Yihu Xu

Format Sitasi

APA MLA BibTeX

Meng, X., Liu, X., Xu, Y., Wu, Y., Li, H., Kim, K. et al. (2025). A Multi-Time-Frequency Feature Fusion Approach for Marine Mammal Sound Recognition. https://doi.org/10.3390/jmse13061101

Akses Cepat

PDF tidak tersedia langsung

Cek di sumber asli →

Lihat di Sumber doi.org/10.3390/jmse13061101

Informasi Jurnal

Tahun Terbit: 2025
Sumber Database: DOAJ
DOI: 10.3390/jmse13061101
Akses: Open Access ✓