aiqbench analysis
This research introduces Spatial FiLM (SpFiLM), a novel conditioning layer that applies voxel-wise scale and shift modulations to improve brain parcellation accuracy on both pre- and post-contrast MRI scans. By addressing the limitations of uniform modulation in standard Feature-wise Linear Modulation, the method achieves a significant relative improvement in Dice scores for clinical workflows relying solely on contrast-enhanced imaging. The work represents a meaningful advancement in adapting deep learning architectures to handle distinct input modalities within medical imaging without increasing model complexity.
Spatial Feature-wise Linear Modulation (SpFiLM) for Contrast Agent-Aware Brain Parcellation
arXiv:2609.07718v1 Announce Type: cross Abstract: Most automated brain parcellation tools are developed and validated on T1-weighted (T1w) MRI. Yet, some clinical workflows for which parcellation is relevant only use contrast-enhanced T1w (T1ce) MRI, on which T1w-trained models are less accurate. We present a unified network that parcellates both pre- and post-contrast agent T1w MRI reliably, trained on a combination of the two with conditioning that spatially modulates its response differently for each. Feature-wise Linear Modulation (FiLM) is a known approach for input-based modulation in networks. It applies a per-channel scale and shift uniformly across the input. However, the appearance change between pre- and post-contrast varies locally across the brain, making FiLM suboptimal for our use case. In this work, we introduce Spatial FiLM (SpFiLM), a conditioning layer whose modulation varies spatially, assembling a voxel-wise scale and shift from image-derived spatial patterns. Using a cohort of 134 patients with paired T1w and T1ce MRI parcellated into 106 classes, the addition of SpFiLM layers in a UNet increased the mean Dice on the test set of 25 patients from 80.2% to 84.1%, a 4.9% relative improvement. Adding SpFiLM layers led to the best performance on both pre- and post-contrast MRI, even when controlling for network parameter counts.
Authors:Pushpendra Singh (1), Joshua R. Astley (1), Roman Rodionov (2), John Duncan (2), Tom Vercauteren (1), Rachel Sparks (1) ((1) School of Biomedical Engineering and Imaging Sciences, King's College London, London, UK, (2) Department of Epilepsy, Queen Square Institute of Neurology, University College London, London, UK)
Abstract:Most automated brain parcellation tools are developed and validated on T1-weighted (T1w) MRI. Yet, some clinical workflows for which parcellation is relevant only use contrast-enhanced T1w (T1ce) MRI, on which T1w-trained models are less accurate. We present a unified network that parcellates both pre- and post-contrast agent T1w MRI reliably, trained on a combination of the two with conditioning that spatially modulates its response differently for each. Feature-wise Linear Modulation (FiLM) is a known approach for input-based modulation in networks. It applies a per-channel scale and shift uniformly across the input. However, the appearance change between pre- and post-contrast varies locally across the brain, making FiLM suboptimal for our use case. In this work, we introduce Spatial FiLM (SpFiLM), a conditioning layer whose modulation varies spatially, assembling a voxel-wise scale and shift from image-derived spatial patterns. Using a cohort of 134 patients with paired T1w and T1ce MRI parcellated into 106 classes, the addition of SpFiLM layers in a UNet increased the mean Dice on the test set of 25 patients from 80.2% to 84.1%, a 4.9% relative improvement. Adding SpFiLM layers led to the best performance on both pre- and post-contrast MRI, even when controlling for network parameter counts.
10 pages, 2 figures. Accepted at MLCN 2026 (MICCAI workshop)
Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
arXiv:2609.07718 [eess.IV]
(or arXiv:2609.07718v1 [eess.IV] for this version)
https://doi.org/10.48550/arXiv.2609.07718 arXiv-issued DOI via DataCite (pending registration)
Full text under the publisher's open licence · 285 words
Related coverage
Game Design Creativity evals
LessWrong · Oct 12, 2026
Graph-MambaNav: Spatial-Temporal Graph Mamba Leveraging Object-Relation Knowledge for Object-Goal Navigation
arXiv · Aug 17, 2026
Automating coherent long-form video generation
Google Research · Sep 24, 2026
PSCT-Net: Geometry-Aware Pediatric Skull CT Reconstruction via Differentiable Back-Projection and Attention-Guided Refinement
arXiv · Sep 11, 2026