Abstract:Many odontocetes produce whistles that feature characteristic contour shapes in spectrogram representations of their calls. Automatically extracting the time × frequency tracks of whistle contours has numerous subsequent applications, including species classification, identification, and density estimation. Deep-learning-based methods, which train models using analyst-annotated whistles, offer a promising way to reliably extract whistle contours. However, the application of such methods can be limited by the significant amount of time and labor required for analyst annotation. To overcome this challenge, a technique that learns from automatically generated pseudo-labels has been developed. These annotations are less accurate than those generated by human analysts but more cost-effective to generate. It is shown that standard training methods do not learn effective models from these pseudo-labels. An improved loss function designed to compensate for pseudo-label error that significantly increases whistle extraction performance is introduced. The experiments show that the developed technique performs well when trained with pseudo-labels generated by two different algorithms. Models trained with the generated pseudo-labels can extract whistles with an F1-score (the harmonic mean of precision and recall) of 86.31% and 87.2% for the two sets of pseudo-labels that are considered. This performance is competitive with a model trained with 12 539 expert-annotated whistles (F1-score of 87.47%).

Detection method for whistles of bottlenose dolphin(tursiops truncatus) based on spectrogram processing

Automatic Segmentation and Identification of Whistles Produced by Dolphins.

Automatic detection of estuarine dolphin whistles in spectrogram images

Spectral Entropy Based Dolphin Whistle Detection Algorithm and Its Possible Application for Biologically Inspired Communication

Time-frequency filtering method for detecting clicks of bottlenose dolphin (Tursiops truncates)

Unsupervised Beluga Whistle End-Point Detection under Burst Pulse Interferences

Automatic classification of delphinids based on the representative frequencies of whistles

Robust unsupervised Tursiops aduncus whistle-event detection using gammatone multi-channel Savitzky-Golay based whistle enhancement

A Novel Tracker of Adaptive Directional Ridge Separation and Prediction for Detecting Whistles

Automated detection of dolphin whistles with convolutional networks and transfer learning

Dolphin Whistle Signal Enhancement Using Nonconvex Time-Frequency Regularized Overlapping Group Shrinkage in Warm Shallow Water

Automated Classification of Dolphin Whistles Based on the Convolutional Neural Network

Application of Wavelet Transform for the Detection of Cetacean Acoustic Signals

Analyzing and Classifying Fish Vocalization in the East China Sea Coastal Area

The Acoustical Features Analysis of Indo-Pacific Humpback Dolphins’ Whistle and Modeling Research

Application to the Call of Southern Right Whale

Using deep learning to track time × frequency whistle contours of toothed whales without human-annotated training data

Robust unsupervised <i>Tursiops aduncus</i> whistle enhancement based on complete ensembled empirical optimal envelope local mean decomposition with adaptive noise

Automated Classification of Tursiops Aduncus Whistles Based on a Depth-Wise Separable Convolutional Neural Network and Data Augmentation.

An EMD Based Automatic Endpoint Detection Method for Cetacean Vocal Signals

Whale Vocalization Classification Using Feature Extraction With Resonance Sparse Signal Decomposition And Ridge Extraction