[MARMAM] New Publication: 'More than a whistle: Automated detection of marine sound sources with a convolutional neural network"
Dear Colleagues, We are excited to announce a new scientific publication from the University of Southampton: 'More than a whistle: Automated detection of marine sound sources with a convolutional neural network' in the journal Frontier of Marine Science (Special Issue Ocean Observation). The article is open access and can be viewed at: https://www.frontiersin.org/articles/10.3389/fmars.2022.879145 [https://www.frontiersin.org/files/MyHome%20Article%20Library/879145/879145_Thumb_400.jpg]<https://www.frontiersin.org/articles/10.3389/fmars.2022.879145> More than a whistle: Automated detection of marine sound sources with a convolutional neural network<https://www.frontiersin.org/articles/10.3389/fmars.2022.879145> www.frontiersin.org Authors: Ellen L White, Paul White, Jonathon Bull, Denise Risch, Suzanne Beck and Ewan Edwards. Abstract The effective analysis of Passive Acoustic Monitoring (PAM) data has the potential to determine spatial and temporal variations in ecosystem health and species presence if automated detection and classification algorithms are capable of discrimination between marine species and the presence of anthropogenic and environmental noise. Extracting more than a single sound source or call type will enrich our understanding of the interaction between biological, anthropogenic and geophonic soundscape components in the marine environment. Advances in extracting ecologically valuable cues from the marine environment, embedded within the soundscape, are limited by the time required for manual analyses and the accuracy of existing algorithms when applied to large PAM datasets. In this work, a deep learning model is trained for multi-class marine sound source detection using cloud computing to explore its utility for extracting sound sources for use in marine mammal conservation and ecosystem monitoring. A training set is developed comprising existing datasets amalgamated across geographic, temporal and spatial scales, collected across a range of acoustic platforms. Transfer learning is used to fine-tune an open-source state-of-the-art ‘small-scale’ convolutional neural network (CNN) to detect odontocete tonal and broadband call types and vessel noise (from 0 to 48 kHz). The developed CNN architecture uses a custom image input to exploit the differences in temporal and frequency characteristics between each sound source. Each sound source is identified with high accuracy across various test conditions, including variable signal-to-noise-ratio. We evaluate the effect of ambient noise on detector performance, outlining the importance of understanding the variability of the regional soundscape for which it will be deployed. Our work provides a computationally low-cost, efficient framework for mining big marine acoustic data, for information on temporal scales relevant to the management of marine protected areas and the conservation of vulnerable species. Reference for the paper: White, E.L. White, P. Bull, J. Risch, D. Beck, S and Edwards, E, 2022. More than a whistle: Automated detection of marine sound sources with a convolutional neural network. Frontiers of Marine Science (9). DOI=10.3389/fmars.2022.879145. Please feel free to contact the lead author Ellen White on behalf of all authors if you have any questions, Ellen White Post-graduate Research Student University of Southampton School of Ocean and Earth Sciences National Oceanography Centre Southampton SO14 3ZH, UK Contact Information: Email: elw1...@soton.ac.uk Phone: 07715926069 Twitter: @OceansE11en ___ MARMAM mailing list MARMAM@lists.uvic.ca https://lists.uvic.ca/mailman/listinfo/marmam
[MARMAM] New Publication: One size fits all? Adaptation of trained CNNs to new marine acoustic environments
Dear Colleagues We are excited to share our new research paper 'One size fits all? Adaptation of trained CNNs to new marine acoustic environments' in the journal Ecological Informatics. The article is open access and can be found at : https://doi.org/10.1016/j.ecoinf.2023.102363One size fits all? Adaptation of trained CNNs to new marine acoustic environments The authors of this work are: Ellen L. White, Holger Klinck, Jonathan M. Bull, Paul R. White & Denise Risch. This article looks at deploying a broadband multi-sound source detector within new soundscapes, identifying how much data we should use when fine-tuning to a new acoustic feild. We hope you find the work interesting. Please feel free to contact Ellen White (elw1...@soton.ac.uk) for any questions or enquiries. Abstract: Convolutional neural networks (CNNs) have the potential to enable a revolution in bioacoustics, allowing robust detection and classification of marine sound sources. As global Passive Acoustic Monitoring (PAM) datasets continue to expand it is critical we improve our confidence in the performance of models across different marine environments, if we are to exploit the full ecological value of information within the data. This work demonstrates the transferability of developed CNN models to new acoustic environments by using a pre-trained model developed for one location (West of Scotland, UK) and deploying it in a distinctly different soundscape (Gulf of Mexico, USA). In this work transfer learning is used to fine-tune an existing open-source ‘small-scale’ CNN, which detects odontocete tonal and broadband call types and vessel noise (operating between 0 and 48 kHz). The CNN is fine-tuned on training sets of differing sizes, from the unseen site, to understand the adaptability of a network to new marine acoustic environments. Fine-tuning with a small sample of site-specific data significantly improves the performance of the CNN in the new environment, across all classes. We demonstrate an improved performance in area-under-curve (AUC) score of 0.30, across four classes by fine-training with only 50 spectrograms per class, with a 5% improvement in accuracy between 50 frames and 500 frames. This work shows that only a small amount of site-specific data is needed to retrain a CNN, enabling researchers to harness the power of existing pre-trained models for their own datasets. The marine bioacoustic domain will benefit from a larger pool of global data for training large deep learning models, but we illustrate in this work that domain adaptation can be improved with limited site-specific exemplars. Reference for the paper: White, E., Klinck, H., Bull, J., White, P. and Risch, D., 2023. One size fits all? Adaptation of trained CNNs to new marine acoustic environments. Ecological Informatics, p.102363. Ellen White Post-graduate Research Student University of Southampton School of Ocean and Earth Sciences National Oceanography Centre Southampton SO14 3ZH, UK Office Location: 164/25 Contact Information: Email: elw1...@soton.ac.uk Phone: 07715926069 ___ MARMAM mailing list MARMAM@lists.uvic.ca https://lists.uvic.ca/mailman/listinfo/marmam
[MARMAM] Request for Harbour Porpoise Broadband Recordings
Hello MARMAM network, I am a post-doc at the University of Southampton, and I am just enquiring if anyone has any broadband Harbour Porpoise acoustic recordings, they would be willing to share with me. The data will be used to test several detections algorithms and may form part of the training data if permissible. I am testing algorithms for their performance in variable soundscapes and ambient noise conditions, but I am currently struggling to source Harbour Porpoise data. Ideally the sample rate would be around 384 kHz, and any quantity of data is suitable. Thanks in advance Please get in contact via: elw1...@soton.ac.uk<mailto:elw1...@soton.ac.uk> All the best Dr Ellen White -- ___ MARMAM mailing list MARMAM@lists.uvic.ca https://lists.uvic.ca/mailman/listinfo/marmam
