Experimental evaluation demonstrates effective real-time separation of ambient ocean noise from ship sound across diverse vessel classes, highlighting viability for marine acoustic monitoring.
This paper presents a single-channel underwater noise separation method that isolates ship-radiated noise from ambient noise. The proposed approach combines Softmax-based time–frequency masking with a stacked LSTM network using residual connections. Experiments conducted under varying input SNR levels (−5 to +10 dB) demonstrate superior separation performance over classical blind source separation techniques and recent deep learning models. Across 11 ship classes, the method achieves the highest SDR and SNR for most categories while maintaining real-time feasibility with a low Real-Time Factor (RTF = 0.02). These results highlight its suitability for practical marine monitoring applications.
No takes yet. Share an insight, caveat, or question.
An et al. (2025) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: