55 citations · 155 across the 13 of their papers we have counts for
8 papers · 1 filter
An Image is Worth Topics: A Visual Structural Topic Model with Pretrained Image Embeddings
Matías Piqueras, Alexandra Segerberg, Matteo Magnani +2
Political scientists are increasingly interested in analyzing visual content at scale. However, the existing computational toolbox is still in need of methods and models attuned to…
Isolated Channel Vision Transformers: From Single-Channel Pretraining to Multi-Channel Finetuning
Wenyi Lian, Patrick Micke, Joakim Lindblad +1
Vision Transformers (ViTs) have achieved remarkable success in standard RGB image processing tasks. However, applying ViTs to multi-channel imaging (MCI) data, e.g., for medical an…
Can representation learning for multimodal image registration be improved by supervision of intermediate layers?
Elisabeth Wetzer, Joakim Lindblad, Nataša Sladoje
Multimodal imaging and correlative analysis typically require image alignment. Contrastive learning can generate representations of multimodal images, reducing the challenging task…
End-to-end Multiple Instance Learning with Gradient Accumulation
Axel Andersson, Nadezhda Koriakina, Nataša Sladoje +1
Being able to learn on weakly labeled data, and provide interpretability, are two of the main reasons why attention-based deep multiple instance learning (ABMIL) methods have becom…
Cross-Modality Sub-Image Retrieval using Contrastive Multimodal Image Representations
Eva Breznik, Elisabeth Wetzer, Joakim Lindblad +1
In tissue characterization and cancer diagnostics, multimodal imaging has emerged as a powerful technique. Thanks to computational advances, large datasets can be exploited to disc…
Fast computation of mutual information in the frequency domain with applications to global multimodal image alignment
Johan Öfverstedt, Joakim Lindblad, Nataša Sladoje
Multimodal image alignment is the process of finding spatial correspondences between images formed by different imaging techniques or under different conditions, to facilitate hete…
INSPIRE: Intensity and spatial information-based deformable image registration
Johan Öfverstedt, Joakim Lindblad, Nataša Sladoje
We present INSPIRE, a top-performing general-purpose method for deformable image registration. INSPIRE brings distance measures which combine intensity and spatial information into…
CoMIR: Contrastive Multimodal Image Representation for Registration
Nicolas Pielawski, Elisabeth Wetzer, Johan Öfverstedt +4
We propose contrastive coding to learn shared, dense image representations, referred to as CoMIRs (Contrastive Multimodal Image Representations). CoMIRs enable the registration of…