Abstract:We present a benchmark for detecting animal vocalizations in marine passive acoustic data including a published dataset, a concrete task and metrics to report. Three deep learning models for bioacoustics are tested on this benchmark and their performances are reported. Future efforts in the field should aim at the reduction of false‐positive rates in order to assure usability of the automated techniques. Passive acoustic monitoring (PAM) is commonly used to obtain year‐round continuous data on marine soundscapes harboring valuable information on species distributions or ecosystem dynamics. This continuously increasing amount of data requires highly efficient automated analysis techniques in order to exploit the full potential of the available data. Here, we propose a benchmark, which consists of a public dataset, a well‐defined task and evaluation procedure to develop and test automated analysis techniques. This benchmark focuses on the special case of detecting animal vocalizations in a real‐world dataset from the marine realm. We believe that such a benchmark is necessary to monitor the progress in the development of new detection algorithms in the field of marine bioacoustics. We ultimately use the proposed benchmark to test three detection approaches, namely ANIMAL‐SPOT, Koogu and a simple custom sequential convolutional neural network (CNN), and report performances. We report the performance of the three detection approaches in a blocked cross‐validation fashion with 11 site‐year blocks for a multi‐species detection scenario in a large marine passive acoustic dataset. Performance was measured with three simple metrics (i.e., true classification rate, noise misclassification rate and call misclassification rate) and one combined fitness metric, which allocates more weight to the minimization of false positives created by noise. Overall, ANIMAL‐SPOT performed the best with an average fitness metric of 0.6, followed by the custom CNN with an average fitness metric of 0.57 and finally Koogu with an average fitness metric of 0.42. The presented benchmark is an important step to advance in the automatic processing of the continuously growing amount of PAM data that are collected throughout the world's oceans. To ultimately achieve usability of developed algorithms, the focus of future work should be laid on the reduction of the false positives created by noise.

Simulation of Marine Mammal Calls in Deep-Sea Environment

Bio-Inspired Steganography for Secure Underwater Acoustic Communications

Adaptive Modulation and Coding for Underwater Acoustic Communication Based on Neural Networks

WhaleNet: a Novel Deep Learning Architecture for Marine Mammals Vocalizations on Watkins Marine Mammal Sound Database

Automated Call Detection for Acoustic Surveys with Structured Calls of Varying Length

A classification method of marine mammal calls based on two-channel fusion network

Towards the topology of autoencoder of calls versus clicks of marine mammal

Data Enhancement of Underwater High-Speed Vehicle Echo Signals Based on Improved Generative Adversarial Networks

Classification on Underwater Acoustic Propagation Model U Sing Convolutional Neural Network

Automatic Modulation Classification for Underwater Acoustic Communication Signals Based on Deep Complex Networks

Development of deep neural networks for marine mammal call detection using an open-source, user friendly tool

Marine Mammal Species Classification Using Convolutional Neural Networks and a Novel Acoustic Representation

DCC-GAN-based channel emulator for underwater wireless optical communication systems

Feature extraction and classification of deep-sea mobile underwater acoustic channels

The Effects of Signal-to-Noise Ratio on Generative Adversarial Networks Applied to Marine Bioacoustic Data

Underwater Acoustic Target Recognition Based on Data Augmentation and Residual CNN

Low-Resource Generation Method for Few-Shot Dolphin Whistle Signal Based on Generative Adversarial Network

Deep learning in marine bioacoustics: a benchmark for baleen whale detection

Application to the Call of Southern Right Whale

Bidirectional Generative Adversarial Representation Learning for Natural Stimulus Synthesis

WhistleGAN for Biomimetic Underwater Acoustic Covert Communication