Sara Björk defending her PhD work
Image:
Harald Lykke

Sara Björk defending her PhD work

Successful PhD Defence by Sara Björk

Sara Marie Björk defended her PhD thesis “Deep convolutional regression modelling for forest parameter retrieval” on October 6th 2023 at UiT The Arctic University of Norway.

Solid PhD defense at the Visual Intelligence Research Centre

We proudly congratulate Sara Maria Björk with her excellent defense of her PhD thesis “Deep convolutional regression modelling for forest parameter retrieval” on October 6th 2023 at UiT The Arctic University of Norway. Sara Börk was affiliated with the Visual Intelligence research center during her PhD work, and is now a KSAT employee.

Summary of thesis:

Accurate forest monitoring is crucial as forests are major global carbon sinks. Additionally, accurate prediction of forest parameters, such as forest biomass and stem volume (SV), has economic importance. Therefore, the development of regression models for forest parameter retrieval is essential.

Existing forest parameter estimation methods use regression models that establish pixel-wise relationships between ground reference data and corresponding pixels in remote sensing (RS) images. However, these models often overlook spatial contextual relationships among neighbouring pixels, limiting the potential for improved forest monitoring. The emergence of deep convolutional neural networks (CNNs) provides opportunities for enhanced forest parameter retrieval through their convolutional filters that allow for contextual modelling. However, utilising deep CNNs for regression presents its challenges. One significant challenge is that the training of CNNs typically requires continuous data layers for both predictor and response variables. While RS data is continuous, the ground reference data is sparse and scattered across large areas due to the challenges and costs associated with in situ data collection.

This thesis tackles challenges related to using CNNs for regression by introducing novel deep learning-based solutions across diverse forest types and parameters. To address the sparsity of available reference data, RS-derived prediction maps can be used as auxiliary data to train the CNN-based regression models. This is addressed through two different approaches.

Regression U-Net architecture used for image-to-image translation between Sentinel-1 (SAR) input and Lidar-derived and ground reference data of stem volume as pseudo-targets and targets

Above: Illustration on how conventional statistical or ML-based regression models (f in the image) perform regression between single pixels of Lidar measurements and a sparse set of ground reference measurements of AGB (biomass) or forest stem volume (SV). As these models considers each pixel in the input data individually, these models are non-contextual. A Lidar-derived AGB or SV prediction map can be created by use of the regression model f.
Below: Illustration of how a CNN-based regression model (g) utilise the neigbourhood of pixels through the convolutional filters in the learning of the realionship between input data of Sentine-1 (SAR scenes) and a lidar-derived prediction map as a target. Due to the convolutional filters, the CNN-based regression models are contextual

Dean Arne Smalås, Professor Anthony Paul Doulgeries, Dr. Sara Björk, Dr. Oleg Antropov, Professor Stian Normann Anfinsen and Dr. Alba Ordonez (Back)

Evaluation Committee:

  • Dr. Oleg Antropov, senior researcher at VTT Technical Research Center of Finland (1. Opponent)
  • Dr. Alba Ordonez, senior researcher at Norwegian Computing Center (2. Opponent)
  • Professor Anthony Paul Doulgeries, dept. of Physics and Technology, UiT (internal member and committee leader)

Supervisors:

  • Professor Stian Normann Anfinsen, dept. of Physics and Technology, UiT (main supervisor)
  • Professor Robert Jenssen, dept. of Physics and Technology, UiT

The Master of ceremony was Professor Arne Smalås, Dean of the Faculty of Science and Technology, UiT.

Download the thesis HERE

Latest news

My Research Stay at Visual Intelligence: María Castro Fernández

July 24, 2024

María Castro Fernández is a PhD Candidate at the Research Institute of Applied Microelectronics at Universidad de Las Palmas de Gran Canaria, in Spain. She visited Visual Intelligence in Tromsø from January to June 2024.

Visual Intelligence at Arendalsuka 2024: Hvordan implementerer vi KI for bruk i helsesektoren på en trygg måte?

July 22, 2024

AI tools wield enormous potential within the health sector. How do we ensure that such tools are safely implemented and used within this field? The event will take place on August 14th, 11:45-12:45 at Bankgården, Strandgaten 2-4.

My Research Stay at Visual Intelligence: Thea Brüsch

July 1, 2024

Thea Brüsch is a PhD student at the Section for Cognitive Systems at the Technical University of Denmark (DTU). She visited the UiT Machine Learning Group from January to June 2024.

Teknisk Ukeblad: Finner leverkreft med kunstig intelligens

June 26, 2024

Analyse av CT-bilder er sentralt for å avdekke mulig kreft i lever, men tar ofte mye tid. Forskere ved SFI Visual Intelligence har nå utviklet et kunstig intelligens-system som på sikt kan bidra til å effektivisere analysearbeidet (News story in Norwegian).

Visual Intelligence paper featured in EurekAlert

June 24, 2024

We are happy to announce that a Visual Intelligence paper "The 3-billion fossil question: How to automate classification of microfossils" was featured in the scientific newspaper EurekAlert!

Successful course on generative AI at UiT, in collaboration with Visual Intelligence

June 21, 2024

Visual Intelligence organized a special curricular course on Generative AI at UiT The Arctic University of Norway this spring. The course covered Large Language Models, Generative AI for images, and Generative AI for audio.

Become an NLDL sponsor!

June 20, 2024

Our sponsorship packages provide a number of benefits, such as registration and networking benefits for company employees, company and brand visibility, and more.

NRK Radio: UiT and UNN develop AI to detect liver cancer

June 18, 2024

Associate professor and principal investigator Kristoffer Wickstrøm recently talked about a collaborative Visual Intelligence project between UiT and UNN on NRK Radio. The project aims to develop AI algorithms to detect liver cancer on CT images.

Successful PhD defense by Tollef Jahren

June 14, 2024

We congratulate Tollef Jahren for successfully defending his PhD thesis "Deep learning for detecting valvular events and suppressing reverberations in cardiac ultrasound" for the degree Philosophiae Doctor at Department of Informatics, University of Oslo.

Exchange stays and pizza on the menu at VIGS social event in Tromsø

June 4, 2024

The Visual Intelligence Graduate School (VIGS) recently gathered our Tromsø based researchers at Árdna - The Sami Cultural Center for a social event about exchange stays.

Norsk regnesentral podcast: «In the 90s, everybody knew that AI didn't work.» Hvorfor hadde Sergey Brin rett da han sa det?

May 15, 2024

Listen to VI co-director Line Eikvil and research director Anders Løland (Norsk Regnesentral) talk about developments before and after the breakthrough of deep learning in the NR podcast, Sannsynligvis VIKTIG (Norwegian dialogue).