An integrated underwater acoustic pipeline combining U-Net denoising, ResNet classification, and autoencoder novelty detection reports 71% classification and 91% novelty detection on Deepship data, but the novelty metric is below a trivial baseline.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SD 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
AquaSignal: An Integrated Framework for Robust Underwater Acoustic Analysis
An integrated underwater acoustic pipeline combining U-Net denoising, ResNet classification, and autoencoder novelty detection reports 71% classification and 91% novelty detection on Deepship data, but the novelty metric is below a trivial baseline.