ORCA-SPOT: An Automatic Killer Whale Sound Detection Toolkit Using Deep Learning

Abstract

Large bioacoustic archives of wild animals are an important source to identify reappearing communication patterns, which can then be related to recurring behavioral patterns to advance the current understanding of intra-specific communication of non-human animals. A main challenge remains that most large-scale bioacoustic archives contain only a small percentage of animal vocalizations and a large amount of environmental noise, which makes it extremely difficult to manually retrieve sufficient vocalizations for further analysis – particularly important for species with advanced social systems and complex vocalizations. In this study deep neural networks were trained on 11,509 killer whale ( Orcinus orca ) signals and 34,848 noise segments. The resulting toolkit ORCA-SPOT was tested on a large-scale bioacoustic repository – the Orchive – comprising roughly 19,000 hours of killer whale underwater recordings. An automated segmentation of the entire Orchive recordings (about 2.2 years) took approximately 8 days. It achieved a time-based precision or positive-predictive-value (PPV) of 93.2% and an area-under-the-curve (AUC) of 0.9523. This approach enables an automated annotation procedure of large bioacoustics databases to extract killer whale sounds, which are essential for subsequent identification of significant communication patterns. The code will be publicly available in October 2019 to support the application of deep learning to bioaoucstic research. ORCA-SPOT can be adapted to other animal species.

Evidences of Intra-Group Orca Call Rate Modulation Using A Small-Aperture Four Hydrophone Array

“…Our shore based study looked at orca communication in the wild at a unique level of precision. Other shore based fixed hydrophone systems enabled the detection and tracking of orcas from the same population [36][37][38] but did not succeed in associating individuals with calls, or determine precise individual pattern variations of communication between groups. This work does not pretend to provide all explanations for the type of calls, but rather the CR.…”

EXACT: a collaboration toolset for algorithm-aided annotation of images with annotation version control

“…While the EXACT platform is primarily developed for cooperative interdisciplinary research on microscopy images, its flexibility extends to other research areas without adaptation. We therefore showcase its use in a project that aims at deepening the understanding of killer whales (Orcinus Orca) and their large variety of different sound types 20 . In this study, EXACT is used to cluster and visualise the spectral shape of machine-pre-segmented killer whale audio samples (Fig.…”

Deep Learning and Domain Transfer for Orca Vocalization Detection

“…A database called Orchive [13] was built in collaboration with this laboratory, aggregating manually segmented orca calls from 1980 up till today. It is the only large scale open corpus of orca's acoustic emissions, and can be used to train a deep learning model [3]. However, the fact that this database consists only of manually selected calls induces a bias (e.g.…”

“…The results presented in V display mean ± standard deviation of the 10 scores. To measure the classification performances of the models, their output being a real value, (5) 1x1x32x55125 Conv2D (3,5) 1x32x32x27563 MaxPool 1x32x32x13781 Conv3D (3,3,5) 8x16x16x3446 Conv3D (3,3,5) 32x8x8x862 Conv3D (3,3,5) 64x4x4x431 Conv3D (2,2,5) 128x3x3x216 Conv3D (1,1,1 a discrimination threshold needs to be chosen. To emphasis the domain transfer problematic, thresholds were chosen to optimize the score on the training set (threshold for which the true positive rate equals the true negative rate).…”

“…The question of the tranferability of the learned models to different acoustic conditions is still relevant, we will try to approach it in this paper. For this task of orca vocalization detection using deep learning, certainly the main paper in the literature is Orca-spot [3]. In this paper, several sources of data have been used.…”