Deep Learning-Based Identification of Arctic Ocean Boundaries and Near-Surface Phenomena in Underwater Echograms
Bibliographic record
Abstract
Monitoring marine environments is a crucial part of understanding the impact of oceans on global climate and their importance for biodiversity and ecological systems, particularly in the Arctic region. Underwater active acoustic surveys with moored multi-frequency echosounders allow for the continuous collection of valuable data reflecting the complex dynamics of these environments. This paper addresses the automatic identification of sea surface boundaries and near-surface phenomena in echograms using deep learning methods to support researchers such as biologists in their work, who typically rely on time-consuming manual analyses. We propose a two-step process that first characterizes echograms according to the surface conditions using an image classification paradigm and then identifies the sea surface boundary and near-surface bubbles and their extent in the water column using a semantic segmentation paradigm. Segmentation is carried out using surface type-specific models, which perform better than a single global segmentation model. We also propose learning strategies, such as a custom boundary loss function, that further improve performance. Experiments with various image classification and semantic segmentation architectures allow us to select the most efficient models for Arctic echogram analysis that, when used in conjunction within our proposed pipeline and our learning strategies, offer excellent results.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".