ONCship: A Large-Scale Underwater Acoustic Corpus for Ship Recognition
Recent advances in deep learning have significantly improved underwater acoustic vessel recognition.
ProofPaper ↗
Key points
- To address this limitation, we introduce ONCship, a large-scale underwater acoustic dataset comprising 4521 hours of curated recordings from Ocean Networks Canada (ONC) hydrophones and Automatic Identification System (AIS) data.
- Experiments on DeepShip and ShipsEar demonstrate the effectiveness of ONCship for acoustic representation learning, with joint pre-training on ONCship and AudioSet achieving a classification accuracy of 62.45% on DeepShip through linear probing.
- Furthermore, incorporating ONCship as auxiliary training data improves downstream vessel-type classification performance.
- Overall, ONCship provides a valuable resource for underwater acoustic representation learning and vessel recognition, while its rich vessel-level metadata enables further exploration of MMSI-Level Vessel Classification.
Sources (1)
- [1]ONCship: A Large-Scale Underwater Acoustic Corpus for Ship RecognitionarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 14, 07:48 AM
Recent advances in deep learning have significantly improved underwater acoustic vessel recognition.
To address this limitation, we introduce ONCship, a large-scale underwater acoustic dataset comprising 4521 hours of curated recordings from Ocean Networks Canada (ONC) hydrophones and Automatic Identification System (AIS) data.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 14, 2026Asymptotic Behavior of Transformers
- Oct 14, 2026Player-1: Learning to Play from Synthetic Experience
- Oct 13, 2026pydantic/pydantic-ai v2.56.0: v2.56.0 (2026-10-13)
- Oct 13, 2026Bytes to Threats: Byte-Sequence Transformer For Malware Detection
- Oct 13, 2026OmniCompass: A Holistic Benchmark for Omni-Modal Video Grounding and Understanding
- Oct 13, 20264LC: 4D Latent Cache for Efficient Conditioning in Dynamic World Modeling