Article
Computer Science, Artificial Intelligence
David Sosa-Trejo, Antonio Bandera, Martin Gonzalez, Santiago Hernandez-Leon
Summary: Since the 19th century, scientists have tried to quantify species distributions using techniques such as direct counting and microscopes. Automatic image processing and classification methods are now being utilized to avoid manual procedures for classifying marine plankton. This article summarizes the techniques proposed for classifying marine plankton from the beginning of this field to the present day, focusing on automatic methods that utilize image processing.
ARTIFICIAL INTELLIGENCE REVIEW
(2023)
Article
Engineering, Electrical & Electronic
Baogui Sun, Xinwei Luo
Summary: Underwater acoustic target recognition (UATR) technology based on deep learning and automatic encoding has emerged as an important research direction in recent years. However, the existing methods lack self-adaptability for different data due to the complex and changeable underwater environment, leading to unsatisfactory recognition outcomes. This study introduces the concept of contrastive learning into UATR and proposes a model named Contrastive Coding for UATR (CCU). The CCU model, based on unsupervised contrastive learning framework, effectively generates adaptable automatic features for different underwater acoustic data, achieving excellent recognition performance compared to other automatic encoding models.
IET RADAR SONAR AND NAVIGATION
(2023)
Article
Computer Science, Information Systems
Alberto Testolin, Dror Kipnis, Roee Diamant
Summary: Accurate detection and quantification of submerged targets has always been a challenging task in marine exploration. This study presents a deep learning approach using reflections to detect the pattern of moving fish. Training the network on real reflections with data augmentation techniques shows a more favorable precision-recall trade-off.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2022)
Article
Energy & Fuels
Sihan Yang, Qiguo Liu, Xiaoping Li, Youjie Xu
Summary: This study proposes an automated framework for well test model identification using syntactic pattern recognition. The framework consists of six steps and can effectively address the non-uniqueness of different reservoir models. The findings of this study are important for better understanding the process of well test experts in completing the task of model identification.
PETROLEUM SCIENCE AND TECHNOLOGY
(2022)
Article
Environmental Sciences
Zeshu Yu, Marty Kwok-Shing Wong, Jun Inoue, Sk Istiaque Ahmed, Tomihiko Higuchi, Susumu Hyodo, Sachihiko Itoh, Kosei Komatsu, Hiroaki Saito, Shin-ichi Ito
Summary: Using environmental DNA monitoring, we were able to reveal the spatial distribution patterns of small pelagic fishes in the Kuroshio Current system and hypothesize that predator-prey relationships influence their distribution in these fish communities.
FRONTIERS IN MARINE SCIENCE
(2023)
Article
Ecology
Rune Vabo, Endre Moen, Szymon Smolinski, Ase Husebo, Nils Olav Handegard, Ketil Malde
Summary: The study applied Convolutional Neural Networks with transfer learning on a dataset of Atlantic salmon scales, achieving high accuracy in predicting fish origin, spawning history, and sea age, but lower accuracy for river age. Comparison with human expert readers showed higher agreement in sea age prediction and lower agreement in river age, indicating the difficulty of the task.
ECOLOGICAL INFORMATICS
(2021)
Article
Psychology, Multidisciplinary
Yuseok Jeong, Sang Hee Kim
Summary: The study found that promotion training can reduce the association between loneliness and recognition of anger and fear, while increasing the likelihood of positive evaluation towards surprise expressions. This positive evaluation bias towards surprise faces is strengthened in lonely participants with greater promotion learning.
COMPUTERS IN HUMAN BEHAVIOR
(2021)
Review
Fisheries
Ryan T. Munnelly, Jose C. Castillo, Nils Olav Handegard, Matthew E. Kimball, Kevin M. Boswell, Guillaume Rieucau
Summary: Behavioral information is crucial for understanding the ecosystem interactions of aquatic animals. However, obtaining such information is challenging. Advancements in imaging sonar technology provide new possibilities for studying aquatic animal behavior.
ICES JOURNAL OF MARINE SCIENCE
(2023)
Article
Computer Science, Artificial Intelligence
Chung-Feng Jeffrey Kuo, Jagadish Barman, Chun-Chia Huang
Summary: In this study, repeating patterns on printed fabric were extracted and analyzed using various methods. The results showed the feasibility of using fast Fourier transform to establish a database of repeating patterns. The average time for automatic pattern extraction was 10 seconds per image. The developed similarity analysis had high accuracy, sensitivity, and specificity, making it suitable for analyzing all printing patterns.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Telecommunications
Muhammad Ather Iqbal, Zhijie Wang, Zain Anwar Ali, Shazia Riaz
Summary: An automated system for identification and classification of fish species was introduced in this paper, based on deep convolutional neural networks. The proposed model, a reduced version of the AlexNet model, achieved higher testing accuracy compared to the original AlexNet model.
WIRELESS PERSONAL COMMUNICATIONS
(2021)
Article
Chemistry, Analytical
Vincenzo Caro Fuentes, Ariel Torres, Danny Luarte, Jorge E. Pezoa, Sebastian E. Godoy, Sergio N. Torres, Mauricio A. Urbina
Summary: This study proposes and implements an automated fish landing control system that can identify and classify different species in real time, considerably outperforming the current manual procedure.
Review
Computer Science, Artificial Intelligence
Shuo Meng, Ruru Pan, Weidong Gao, Benchao Yan, Yangyang Peng
Summary: This paper provides a comprehensive review of recent research on automatic recognition of woven fabric structural parameters, highlighting the drawbacks of manual operations based on human eyes and experiences and the advantages of computer-vision-based automatic methods. It offers insights for researchers in the textile industry to understand and utilize automated methods effectively.
ARTIFICIAL INTELLIGENCE REVIEW
(2022)
Article
Engineering, Civil
David P. Williams
Summary: By using tiny CNNs with significantly fewer parameters, this study demonstrates classification performance that matches or even surpasses that of human domain experts. The research represents the first large-scale classification study in the sonar domain, showing that CNNs have strong generalization ability in challenging environments, which will significantly impact their utilization in the underwater remote-sensing community.
IEEE JOURNAL OF OCEANIC ENGINEERING
(2021)
Article
Biodiversity Conservation
Kui Zhang, Miao Li, Jiajun Li, Mingshuai Sun, Youwei Xu, Yancong Cai, Zuozhi Chen, Yongsong Qiu
Summary: Marine ecosystems are facing the combined impacts of overfishing and climate change, making it crucial to understand how fish stocks respond to climate change for ecosystem-based fisheries management. Research showed dramatic fluctuations in small pelagic fish populations in the Beibu Gulf after La Niña events, creating uncertainty for the marine ecosystem.
ECOLOGICAL INDICATORS
(2022)
Article
Environmental Sciences
Megan F. McKenna, Simone Baumann-Pickering, Annebelle C. M. Kok, William K. Oestreich, Jeffrey D. Adams, Jack Barkowski, Kurt M. Fristrup, Jeremy A. Goldbogen, John Joseph, Ella B. Kim, Anke Kugler, Marc O. Lammers, Tetyana Margolina, Lindsey E. Peavey Reeves, Timothy J. Rowell, Jenni A. Stanley, Alison K. Stimpert, Eden J. Zang, Brandon L. Southall, Carrie C. Wall, Sofie Van Parijs, Leila T. Hatch
Summary: Soundscapes provide rich descriptions of composite acoustic environments, but characterizing marine soundscapes simply through sound levels is incomplete. Sources contributing to sound levels shift with changes in biological patterns, physical forces, and human activity, sometimes influencing each other. Integrated methods are needed to interpret soundscapes effectively in marine resource management.
FRONTIERS IN MARINE SCIENCE
(2021)
Article
Fisheries
Roland Proud, Nils Olav Handegard, Rudy J. Kloser, Martin J. Cox, Andrew S. Brierley
ICES JOURNAL OF MARINE SCIENCE
(2019)
Article
Fisheries
Heriberto A. Garcia, Chenyang Zhu, Matthew E. Schinault, Anna I. Kaplan, Nils Olav Handegard, Olav Rune Godo, Heidi Ahonen, Nicholas C. Makris, Delin Wang, Wei Huang, Purnima Ratilal
ICES JOURNAL OF MARINE SCIENCE
(2019)
Article
Fisheries
Vaneeda Allken, Nils Olav Handegard, Shale Rosen, Tiffanie Schreyeck, Thomas Mahiout, Ketil Malde
ICES JOURNAL OF MARINE SCIENCE
(2019)
Article
Multidisciplinary Sciences
Neil Anders, Kirsten Howarth, Bjorn Totland, Nils Olav Handegard, Maria Tenningen, Michael Breen
Article
Fisheries
Maria Tenningen, Armin Pobitzer, Nils Olav Handegard, Karen de Jong
ICES JOURNAL OF MARINE SCIENCE
(2019)
Article
Geosciences, Multidisciplinary
Savannah Carolyn Myers, Anders Thorsen, Szymon Smolinski, Jane Aanestad Godiksen, Ketil Malde, Nils Olav Handegard
GEOSCIENCE DATA JOURNAL
(2020)
Article
Fisheries
Arne Johannes Holmin, Erik A. Mousing, Solfrid S. Hjollo, Morten D. Skogen, Geir Huse, Nils Olav Handegard
ICES JOURNAL OF MARINE SCIENCE
(2020)
Article
Geosciences, Multidisciplinary
Vaneeda Allken, Shale Rosen, Nils Olav Handegard, Ketil Malde
Summary: Developing high-performing machine learning algorithms requires large amounts of annotated data, which can be costly and labor-intensive to manually annotate. Researchers have provided a curated set of fish image data with necessary software tools for generating synthetic images and annotations, facilitating the testing of classifier performance.
GEOSCIENCE DATA JOURNAL
(2021)
Article
Fisheries
Changkyu Choi, Michael Kampffmeyer, Nils Olav Handegard, Arnt-Borre Salberg, Olav Brautaset, Line Eikvil, Robert Jenssen
Summary: In this study, a novel semi-supervised deep learning method is proposed for acoustic target classification in multi-frequency echosounder data. The method leverages both annotated and unannotated data samples in one model by optimizing two inter-connected objectives: a clustering objective and a classification objective. The proposed method achieves higher accuracy compared to conventional semi-supervised and fully supervised deep learning methods in the evaluation using echosounder data from the sandeel case study in the North Sea.
ICES JOURNAL OF MARINE SCIENCE
(2021)
Article
Fisheries
Vaneeda Allken, Shale Rosen, Nils Olav Handegard, Ketil Malde, David Demer
Summary: By training a deep learning algorithm on images collected from a trawl-mounted camera system, this study has successfully developed an automated fish detection and counting system, achieving a high average precision on a test set. The system has the potential to be integrated into regular trawl surveys for efficient and accurate fish monitoring.
ICES JOURNAL OF MARINE SCIENCE
(2021)
Article
Engineering, Civil
Changkyu Choi, Michael Kampffmeyer, Nils Olav Handegard, Arnt-Borre Salberg, Robert Jenssen
Summary: Multifrequency echosounder data can provide a comprehensive understanding of the underwater environment without causing disturbance. The analysis of this data is crucial for the marine ecosystem, and semantic segmentation has gained attention in the fisheries and aquatic industry due to its potential in estimating marine organism abundance. However, the reliance on annotated training data is a major challenge, as it is expensive and time-consuming. In response, a novel approach is proposed that combines supervised and unsupervised learning to develop a data-efficient and accurate semisupervised semantic segmentation method.
IEEE JOURNAL OF OCEANIC ENGINEERING
(2023)
Article
Fisheries
K. McQueen, J. E. Skjaeraasen, D. Nyqvist, E. M. Olsen, O. Karlsen, J. J. Meager, P. H. Kvadsheim, N. O. Handegard, T. N. Forland, K. de Jong, L. D. Sivle
Summary: The fine-scale behavioural responses of Atlantic cod to airgun exposure over an extended period were investigated using an acoustic telemetry positioning system on a spawning ground in Norway. The results suggest that relatively distant seismic surveys do not substantially alter cod behaviour during the spawning period at received sound exposure levels varying between 115 and 145 dB re 1 mu Pa(2)s over a 5-d period.
ICES JOURNAL OF MARINE SCIENCE
(2023)
Article
Fisheries
Sai Geetha Seri, Matthew Edward Schinault, Seth Michael Penna, Chenyang Zhu, Lise Doksaeter Sivle, Karen de Jong, Nils Olav Handegard, Purnima Ratilal
Summary: In an experiment in spring 2019, a prototype eight-element coherent hydrophone array was deployed to monitor spawning vocalizations of Norwegian coastal cod. The array recorded cod sounds and analysed their time-frequency characteristics and source level distribution. The study also estimated the spatial dependence of received cod vocalization rates using the array measurements, and found significantly reduced vocalization detection regions in shallow-water areas of the experimental site. The towable hydrophone array proved invaluable in providing continuous spatial coverage and complementing fixed sensor systems.
ICES JOURNAL OF MARINE SCIENCE
(2023)
Article
Fisheries
Ahmet Pala, Anna Oleynik, Ingrid Utseth, Nils Olav Handegard
Summary: Acoustic surveys play a crucial role in fisheries management by identifying different types of backscatter signals. However, conventional convolutional neural networks may struggle with imbalanced data. This study proposes a sampling strategy to address this issue and achieves accurate target classification by balancing the training and validation data.
ICES JOURNAL OF MARINE SCIENCE
(2023)
Review
Environmental Sciences
Derrick Snowden, Vardis M. Tsontos, Nils Olav Handegard, Marcos Zarate, Kevin O'Brien, Kenneth S. Casey, Neville Smith, Helge Sagen, Kathleen Bailey, Mirtha N. Lewis, Sean C. Arms
FRONTIERS IN MARINE SCIENCE
(2019)