gabei62

Gabriel Eilertsen

Associate Professor, Docent

My research focuses on machine learning for understanding, manipulation and creation of images, with applications in computer graphics and medical image diagnosis.

Deep learning and images

I have a PhD in computer graphics and image processing, with focus on high dynamic range (HDR) imaging and machine learning. I am currently an AI/MLX Assistant Professor within the Wallenberg AI, Autonomous Systems and Software Program (WASP). I am a core member of the Analytic Imaging Diagnostics Arena (AIDA), which is a national arena for research and innovation around AI for medical imaging diagnostics. I am an affiliated researcher of the Center for Medical Image Science and Visualization (CMIV).

My research is focused on machine learning, and particularly by means of deep learning, for computer vision and image processing. The research covers both applied problems within computer graphics and medical diagnosis, as well as more fundamental problems in image generation and data-centric deep learning. Examples of current research projects include:

Synthetic images for machine learning

We explore different methods for generating synthetic images for training of deep neural networks. This includes both conventional methods within computer graphics as well as deep generative modeling, and different combinations of these two (such as neural rendering). We investigate techniques for data-centric machine learning using synthetic data, for augmentation, anonymization, and testing. I currently have a CENIIT project on generative deep learning for data-centric medical imaging. For information, see the project page.

High dynamic range imaging

I have a background in high dynamic range (HDR) imaging, with projects related to tone-mapping, compression, evaluation, and deep learning-based HDR image reconstruction. Recent work include evaluation of methods for deep single-image HDR reconstruction. We show how unreliable existing evaluation protocols are, and suggest techniques for improved correlation between objective metrics and perceptual experiments.



Publications

2025

Milda Poceviciute, Yifan Ding, Ruben Bromée, Gabriel Eilertsen (2025) Out-of-distribution detection in digital pathology: Do foundation models bring the end to reconstruction-based approaches? Computers in Biology and Medicine, Vol. 184, Article 109327 (Article in journal) Continue to DOI
Jens Nilsson, Jonas Unger, Gabriel Eilertsen (2025) Self-Prioritizing Multi-Agent Reinforcement Learning for Conflict Resolution in Air Traffic Control with Limited Instructions AEROSPACE, Vol. 12, Article 88 (Article in journal) Continue to DOI

2024

Tahereh Dehdarirad, Ericka Johnson, Gabriel Eilertsen, Saghi Hajisharif (2024) Enhancing Tabular GAN Fairness: The Impact of Intersectional Feature Selection
George Baravdish, Gabriel Eilertsen, Rym Jaroudi, Tomas Johansson, Lukáš Malý, Jonas Unger (2024) A Hybrid Sobolev Gradient Method for Learning NODEs Operations Research Forum, Vol. 5, Article 91 (Article in journal) Continue to DOI
Gabriel Eilertsen, Daniel Jönsson, Jonas Unger, Anders Ynnerman (2024) Model-invariant Weight Distribution Descriptors for Visual Exploration of Neural Networks en Masse EuroVis 2024 - Short Papers (Conference paper) Continue to DOI

Organisation

News