Automatic Detection and Visualization of Marine Mammal Sounds in British-Columbia, Canada.
Bibliographic record
Abstract
Planned developments of marine terminals along the British Columbia (B.C.) coast will lead to increased exposures of marine mammals to noise from vessel traffic. Monitoring marine mammals off B.C. is becoming increasingly important for assessing their interactions with anthropogenic activities. Over the last decade, a large number of autonomous and cabled hydrophones have been deployed in B.C. waters to detect marine mammal calls and to measure ambient noise levels. These systems collect large data volumes that require considerable time and effort to fully analyze and interpret. That often precludes a comprehensive analysis, and data are commonly archived before their full value is exploited. In this study we describe techniques to automatically detect, classify, and visualize sounds produced by several species of cetaceans that frequent B.C. waters. Blue and fin whale calls are detected using a spectrogram correlation approach. Killer whale, humpback whale and Pacific white sided dolphin vocalizations are detected by calculating the local variance of intensity in the spectrogram and then classified using a random forest classifier. A web visualization interface (PAMview) is used to easily navigate through, display, and share detection and classification results. This interface is organized as three interconnected visualization panels: 1) a geographic interface displays maps of the total number of detections for each species at all monitoring locations within an adjustable time period, 2) an interactive detection time series displays temporal variations of detections for several species at a given monitoring location, and 3) a multimedia panel allows the user to visualize spectrograms, listen to sounds and to validate automatic detections. A demonstration of the acoustic monitoring system developed will be performed using archival and real-time data from the VENUS and NEPTUNE ocean observatories.
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 | 0.000 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.006 | 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".