Deep Semi-Supervised Semantic Segmentation in Multi-Frequency Echosounder Data
AuthorChoi, Changkyu; Kampffmeyer, Michael; Jenssen, Robert; Handegard, Nils Olav; Salberg, Arnt-Børre
Multi-frequency echosounder data can provide a broad understanding of the underwater environment in a non-invasive manner. The analysis of echosounder data is, hence, a topic of great importance for the marine ecosystem. Semantic segmentation, a deep learning based analysis method predicting the class attribute of each acoustic intensity, has recently been in the spotlight of the fisheries and aquatic industry since its result can be used to estimate the abundance of the marine organisms. However, a fundamental problem with current methods is the massive reliance on the availability of large amounts of annotated training data, which can only be acquired through expensive handcrafted annotation processes, making such approaches unrealistic in practice. As a solution to this challenge, we propose a novel approach, where we leverage a small amount of annotated data (supervised deep learning) and a large amount of readily available unannotated data (unsupervised learning), yielding a new data-efficient and accurate semi-supervised semantic segmentation method, all embodied into a single end-to-end trainable convolutional neural networks architecture. Our method is evaluated on representative data from a sandeel survey in the North Sea conducted by the Norwegian Institute of Marine Research. The rigorous experiments validate that our method achieves comparable results utilizing only 40 percent of the annotated data on which the supervised method is trained, by leveraging unannotated data. The code is available at https://github.com/SFI-Visual-Intelligence/PredKlus-semisup-segmentation.
Is part ofChoi, C. (2023). Advancing Deep Learning for Marine Environment Monitoring. (Doctoral thesis). https://hdl.handle.net/10037/29267.
CitationChoi, Kampffmeyer, Jenssen, Handegard, Salberg. Deep Semi-Supervised Semantic Segmentation in Multi-Frequency Echosounder Data. IEEE Journal of Oceanic Engineering. 2023
MetadataShow full item record
Copyright 2023 The Author(s)
Except where otherwise noted, this item's license is described as Attribution 4.0 International (CC BY 4.0)
Showing items related by title, author, creator and subject.
Nilsen, Henrik Fjeld (Master thesis; Mastergradsoppgave, 2021-05-18)This report discusses the possibility for wireless charging solutions for electric vessels, with a focus on offshore wind turbine service. Where the charging time is minimal and safety for crew is important. Different types of wireless technologies have been studied, where the Inductive power transfer (IPT) is shown to be the preferred technology. Inductive power transfer (IPT) grants a safe ...
Innovation and commercialization potential of university-developed arctic ice-tethered platforms. A case study of research-based technology Agwu, Ukeje Jacob; Logvinovskaya, Anna; Phetchpinkaew, Gorn (Master thesis; Mastergradsoppgave, 2017-05-31)The Arctic is a region which is rapidly opening up for business opportunities. However, research has been abundant here for a long time. With this situation comes the avenue to transition technology used in the region from research to commercial purposes. In this thesis, a business case is developed to show the commercial potential of research-based technology. The case technology was developed by ...
Analysis of potential critical equipment and technical system on a modern PSV. Recommending a method for Troms Offshore Management AS Løvmo, Signy Anita (Master thesis; Mastergradsoppgave, 2016-06-01)This thesis is a part of a master’s degree in Technology and Safety in the High North at the University of Tromsø- The Arctic University of Norway. The thesis has been written during the spring semester of 2016. Safety is a large part of maritime operations and all tools to improve safety and reliability is considered. Even in these days when economy in the oil related industry is worse than ever. ...