CRIMAC will improve and automate the interpretation of data and images from modern broadband acoustics on research vessels and fishing boats by using cruises and experimental field research, artificial intelligence, drones and inspection technology.
We published a comprehensive review of 1000 Publications showcasing ML applications in the field. CRIMAC contributed with the review of active acoustics.
New article: How to obtain clear images from in-trawl cameras near the seabed? A case study from the Barents Sea demersal fishing grounds
I am delighted to share my experience attending the Foundations of Computational Mathematics (FoCM) conference held in Paris, France. As the 9th edition of the conference, it follows a series of successful meetings that began with the idea of FoCM in Park City back in 1995. Subsequent conferences took place in Rio de Janeiro (1997), Oxford (1999), Minneapolis (2002), Santander (2005), Hong Kong (2008), Budapest (2011), Montevideo (2014), and Barcelona (2017).
CRIMAC has available two PhD fellowships in marine acoustics, located at Department of Physics and Technology, University of Bergen (UiB), Bergen, Norway. Both PhD fellows will collaborate closely with the Ecosystem Acoustics group at Institute of Marine Research (IMR). This is a great opportunity to work in a world-leading scientific group, with a collegial and professionally stimulating working environment.
Both positions have now been filled, with start-up in September 2023.
This year the symposium adressed the team: "From Echosounders to the Cloud: Transforming Acoustic Data to Information.”
CRIMAC cruise report: Development of acoustic and optic methods for underwater target classification
The cruise report from the joint survey for CRIMAC and LoVe observatory on "G.O. Sars" November 2022 is now published.
The Geilo Winter School is an annual event that provides researchers with a unique opportunity to learn from lectures, collaborate and exchange ideas and experiences. This year, the 23rd edition of the Geilo Winter School was held from January 22 - January 27, 2023, on the topic of Computational Statistics.
CRIMAC had four contributions for the 45th Scandinavian Symposium on Physical Acoustics.
New paper: Training algorithms for acoustic target classification that do not require large amounts of training data
Here we further develop our model of semisuperserved learning to work on pixel-based segmentation. This work is done in collaboration with the Visual Intelligence SFI. You can read more about it here
Direct measurement of fish size is useful in fish farming, fisheries and marine research, and we have tested how broadband echosounders can be used for this purpose.
Early Career Scientist Conference in Canada and a tour of the Fisheries and Marine Institute of Memorial University
PhD student Taraneh Westergerling reports from her presentation at the the ICES/PICES early career scientist conference.
2021 was the first full year for SFI CRIMAC. Despite the pandemic, the center has been able to progress well. The annual report for 2021 is now published and can be accessed here (pdf).
Different data-preparation strategies suitable for acoustic target classification using a deep learning is tested. You will find more information here:
CRIMAC had five contributions to the ICES Working Group on Fisheries Acoustics Science and Technology (WGFAST)
The cruise report from the joint survey for CRIMAC and the IMR sampling gear project on G.O. Sars November 2021 is now published.
Marine researcher Nils Olav Handegard takes the helm after Egil Ona.
CRIMAC had three contribtions for “The 45th Scandinavian Symposium on Physical Acoustics.”
This essay outlines how we use machine learning in marine science in Norway. CRIMAC is collaborating with Visual Intelligence SFI where the Norwegian Computing centre and the University of Tromsø are partners on using machine learning on the vast amounts of data from modern echo sounders.
In this newsletter, project manager Nils Olav Handegard gives an overview of recent CRIMAC activities.
Deep neural networks require a substantial amount of data for training. We have developed a new method for training deep networks for acoustic target classification with only 10 % of the training data compared to traditional methods. The work is a collaboration between the COGMAR project and the Visual Intelligence SFI.
Norway has a long-standing collaboration between science, technology and fishing industry within “smart fisheries”. CRIMAC builds on this collaboration. Click to read our contribution in this special issue of Journal of Technology.
Norway's Institute of Marine Research to use robot platforms and enhanced digital infrastructure to harvest data from the seas.
This is good news for both fishers and fish stocks.
The CRIMAC centre wants the echo sounder itself to be able to distinguish between herring, mackerel, shrimps and gas emissions. Now scientists have completed their first expedition to learn the “dialect” of herring.