Papers for

medical image processing teams

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Random filter losses improve image reconstruction without tuning

RaFL: An Empirical Study of Random Filter Losses for Image Reconstruction

Abstract: Pixel losses train image reconstruction networks toward a point statistic of the target distribution, often averaging away fine textures under uncertainty. Perceptual losses alleviate this issue using pretrained features, but those features are inherently domain-specific. We study a random filter loss (RaFL) that compares prediction and target through filters with random weights, avoiding both pretrained features and task-specific loss engineering. RaFL combines random convolutions, random Fourier features, and random wavelet filters, automatically balancing their contributions through gradient-norm matching, eliminating manual loss-weight tuning and grid search. We systematically investigate three design choices: filter family, weight resampling schedule, and weight initialization. Based on validation PSNR and stability across random seeds, we select the configuration that uses all three filter families, redraws filters every optimization step, and employs Gaussian initialization. We evaluate RaFL on three 2D and two 3D image reconstruction tasks. RaFL outperforms a pixel loss on three tasks and matches it on two, and demonstrates stable performance across random seeds on all five tasks. RaFL adds less than 2% to the backbone's forward FLOPs and no inference cost, requiring $182\times$ fewer loss FLOPs than a VGG19 perceptual loss. These findings suggest that RaFL is a simple, tuning-free, and domain-agnostic loss objective for 2D and 3D image reconstruction.

Fri 13 NovComputer Vision and Pattern RecognitionMachine Learning
The gist
Image reconstruction programs often lose fine details when they only focus on matching pixels exactly. Using features from pretrained networks can help, but those features are usually made for specific types of images. The authors study a method that compares images using random filters instead, which works across different image types without needing special tuning. This method is fast to run and performs as well or better than standard pixel-based methods on various 2D and 3D image tasks.
Open → 2611.17707v1

Improving spherical surface mapping with new sensitivity and preconditioning methods

Representation-Aware Closed-Surface Parameterization: Exact Sensitivity Analysis and Pole-Aware Planar Preconditioning

Abstract: Surface parameterization is a fundamental task in geometry processing and shape analysis. For topologically closed surfaces, it is common to parameterize them onto a standardized closed domain in $\mathbb{R}^3$, with a prime example being a unit sphere for genus-0 closed surfaces. Many mapping methods use planar projections and perform computations on the plane rather than directly on the 3D surface. However, perturbations introduced before projection or geometric evolution propagate through the computation and may affect map quality, convergence, stopping behavior, and geometric validity. In this work, we formulate representation sensitivity as a branchwise input--output problem, separating smooth propagation through active Jacobians from changes in anchors, stopping rules, and fallback events. Under consistent piecewise-linear assembly, the unmodified south-pole Beltrami correction in the FLASH method for spherical conformal parameterization reduces exactly to source-mesh harmonic extension. On a fixed-anchor branch, the corrected map factors into chart transition, harmonic propagation, and inverse stereographic projection, yielding an explicit sensitivity Jacobian and a finite-triangle conditioning bound. Separately, we introduce a bounded five-parameter affine preconditioner over planar scale, translation, and anisotropy and evaluate candidates on the corrected spherical map. Across 30 genus-0 surfaces, it reduces median polar upper-tail log-condition and chord-angle distortion by $10.17\%$ and $10.97\%$, respectively, with both measures improving on every mesh. We also apply the framework to other spherical mapping methods including the conformalized mean-curvature flow and spherical density-equalizing maps. Altogether, our work highlights the importance of representation selection as a structural and practical component of spherical parameterization and downstream computation.

Sat 31 OctComputational GeometryGraphics
The gist
Mapping the surface of 3D shapes like spheres onto flat maps is tricky because small changes during the process can cause big problems in accuracy and quality. The authors analyze how those changes affect the results and introduce a new way to measure and reduce errors near difficult points called poles. They also develop a method to adjust the flat map parameters, making these spherical map projections more reliable and less distorted. This work helps improve how computers process and analyze round shapes by better handling tricky mathematical details.
Open → 2611.00721v1

Factorized gradients speed up complex shape transformations by thousands of times

Factorized Gradients for Scalable Highly-expressive Parametric Diffeomorphisms

Abstract: Diffeomorphisms provide a flexible, topology-preserving mathematical tool for modeling complex spatial or temporal deformations and are used in various scientific fields. However, making these deformations simultaneously highly expressive and computationally efficient remains challenging. Continuous Piecewise-Affine Based (CPAB) transformations (Freifeld et al., ICCV 2015; IEEE TPAMI 2017) offer an attractive solution by parameterizing a family of diffeomorphisms via continuous velocity fields that are piecewise affine with respect to a chosen tessellation of the domain. Importantly, $d=\dim(θ)$, the dimension of the parameter vector $θ$, depends on the fineness of the tessellation rather than on the data resolution. Despite this compact parameterization, CPAB optimization has long been hindered by the tight coupling between trajectory integration and gradient computation with respect to $θ$. We introduce FG-CPAB, a factorized-gradient formulation for scalable CPAB transformations. By separating integration from parameter-space projection, our method reduces gradient computation complexity from $\mathcal{O}(dTN)$ to $\mathcal{O}(TN+dC)$, where $T$ is the number of integration steps, $N$ is the number of transformed points, and $C$ is the number of tessellation cells. This reformulation yields speedups of up to four orders of magnitude (e.g., $\approx 10^4\times$ in 2D), substantial memory savings (e.g., $\approx 500\times$ reduction in peak GPU memory usage in 2D), and dramatically higher expressiveness. This makes fine-tessellation CPAB practical in 2D and 3D for the first time, enabling highly expressive parametric diffeomorphisms.

Mon 26 OctComputer Vision and Pattern Recognition
The gist
Modeling complex shapes or changes in space usually requires balancing how detailed the model is with how fast it can run. The authors improved a particular kind of mathematical transformation called CPAB, which can bend and twist shapes without tearing them. They developed a new way to calculate the changes (gradients) that allows much faster and less memory-heavy updates even for very detailed transformations. This breakthrough makes it practical to use highly detailed models of shape deformation in 2D and 3D applications.
Open → 2610.34766v1

L0 norm improves brain CT to MR image synthesis accuracy

L0-Norm-Based CycleGAN for Unpaired Brain CT-to-MR Synthesis

Abstract: Computed tomography (CT) and magnetic resonance (MR) imaging provide complementary information for the brain, but acquiring both modalities is not always possible. The cycle-consistent adversarial network (CycleGAN) enables unpaired CT-to-MR synthesis, but because the standard L1-norm cycle?consistency penalty grows linearly with the residual magnitude, the network tends to focus on large residuals. This paper introduces an L0-norm-based cycle?consistency term, in which the L0-norm is approximated by a smooth and bounded function. This approximation not only increases sensitivity to small residuals but also effectively suppresses the influence of large residuals. The proposed model and an otherwise identical L1-norm baseline were trained on 102 images and evaluated on 9 test images. The proposed model achieves a lower mean squared error (0.02746 versus 0.03464, a 20.7% reduction), a higher mean peak signal-to-noise ratio (15.80 versus 14.77 dB), and a higher mean structural similarity index (0.5689 versus 0.5407). It also has a lower maximum MSE and higher minimum PSNR and SSIM than the baseline. These results indicate that modifying the cycle-consistency penalty can improve reconstruction accuracy on this dataset.

Fri 23 OctComputer Vision and Pattern Recognition
The gist
Brain CT and MR images show different details but are not always both available for a patient. The authors improved a computer method that transforms CT images into MR images without paired scans using a special math trick called the L0 norm. This trick helps the method pay more attention to small differences and less to big errors, making the generated MR images closer to real ones. Their approach showed better accuracy compared to the usual method in tests with brain images.
Open → 2610.32491v1

SimCortex v2 builds safer brain surfaces with fewer errors

SimCortex v2: Joint Cortical Surface Reconstruction with Near-Zero Collisions and Self-Intersections

Abstract: Reconstructing cortical WM and pial surfaces from structural magnetic resonance imaging (MRI) is a prerequisite for surface-based neuroanatomical analysis, yet remains challenging because the cortex is thin and tightly folded. Reconstruction methods can produce geometric artifacts such as mesh self-intersections and collisions between cortical surfaces, and although recent deep learning methods have reduced reconstruction time from hours to minutes, these artifacts persist. We propose SimCortex v2, a deep learning framework for simultaneous reconstruction of the left and right WM and pial surfaces from T1-weighted MRI. SimCortex v2 estimates topologically correct initial surfaces from a volumetric segmentation and refines all four jointly using multi-scale stationary velocity fields predicted by a ribbon-conditioned, U-Net-like network. We evaluated SimCortex v2 on 560 cases from 14 cohorts, thirteen of them unseen during training, spanning ages 6-89, healthy and clinical populations, and scanners from three vendors. SimCortex v2 matched the surface-distance accuracy of the strongest baseline (average symmetric surface distance 0.253 mm) while showing no detected inter-surface collision in 92.14% of cases and the lowest self-intersection fraction (0.044%) among learning-based methods, whereas every baseline produced at least one collision in every case. Source code, configuration files, pretrained weights, preprocessed data, and the exact evaluation splits are publicly released.

Mon 5 OctComputer Vision and Pattern Recognition
The gist
Reconstructing the thin, folded outer layer of the brain from MRI scans is tough and often creates errors where surfaces collide or intersect themselves. The authors propose SimCortex v2, a deep learning tool that simultaneously builds four brain surfaces with very few such errors. Tested on hundreds of cases from many sources, SimCortex v2 matches top accuracy while greatly reducing these geometric mistakes. The tool and data are made available to others for use and further development.
Open → 2610.07378v1

Bayesian Anything Model improves computational imaging with fewer steps

BAM! Bayesian Anything Model: a foundation model for generative computational imaging

Abstract: Generative models are transforming Bayesian computational imaging, yet the field still lacks physics-aware foundation models. Current practice falls into two camps. Large foundation image models are deployed as plug-and-play priors with zero-shot approximate likelihood guidance, which introduces significant bias and computational cost. Physics-aware generative models avoid this bias, but each is tied to a specific dataset, task and instrument. We introduce BAM (Bayesian Anything Model), a lightweight foundation model for few-step, physics-aware posterior sampling that generalises robustly to unseen data and tasks, zero-shot or with minimal finetuning. BAM upgrades the operator-conditioned Reconstruct Anything Model (RAM) backbone (Terris et al.) into a conditional flow map, so instrument physics is specified at inference time rather than fixed during training. BAM has just 36M parameters and is pre-trained jointly on large image corpora and libraries of forward operators. A single network then draws posterior samples in a few steps, with no likelihood approximation and no guidance weights to tune. Across linear inverse problems on FFHQ, AFHQ, LSUN, DIV2K and the Kohler camera-shake benchmark, BAM outperforms in just 3 steps both specialised models and leading zero-shot methods in sample quality, at a fraction of their computational cost. BAM gives the community an accessible entry point to generative computational imaging, lowers the economic and environmental cost of training imaging models, and opens a new path for research on physics-aware Bayesian computational imaging. Official page: https://bayesian-anything-model.github.io/

Wed 30 SeptComputer Vision and Pattern Recognition
The gist
Computational imaging tries to reconstruct original images from incomplete or noisy data but often struggles with bias or limited methods. The authors introduce the Bayesian Anything Model (BAM), a versatile tool that uses knowledge of the imaging instruments to better guess the original image. BAM is lightweight, can handle many types of imaging problems, and works quickly with minimal tuning. It outperforms other specialized methods on several common image reconstruction challenges while using less computing power.
Open → 2609.39660v1

Multiscale reduction improves brain MRI use with 2D foundation models

Reduce, Then Encode: Multiscale Volumetric Reduction for 2D Foundation Models in Brain MRI

Abstract: Pretrained 2D foundation models offer a practical alternative to dedicated 3D pretraining for brain structural magnetic resonance imaging (sMRI), but their use on volumetric data requires bridging the mismatch between a 2D encoder and a 3D volume input. Existing methods typically encode slices independently and integrate their features afterwards. We introduce Multiscale Volumetric Reduction (MVR), a reduce-then-encode approach that compresses each anatomical view from (D) slices into (M << D) complementary 2D components before foundation-model encoding. MVR combines an uncentered-PCA base component derived from the original through-plane intensities with residual detail components constructed from multiscale spatial descriptors. The reduction is estimated from the training volumes without diagnostic labels or gradient-based optimization and remains fixed thereafter. The resulting components are independently processed by a shared frozen 2D foundation model and concatenated for linear probing. Under this frozen-encoder setting, MVR achieves strong overall performance across ADNI, OASIS, and ABIDE relative to the evaluated 2D-to-3D adaptation methods and simple input-reduction baselines, while also generalizing strongly from ADNI to AIBL.

Mon 28 SeptComputer Vision and Pattern Recognition
The gist
Brain MRI scans are 3D images, but many advanced image analysis tools are designed for 2D pictures. The authors introduce a way to compress each 3D brain image into a few 2D pieces while keeping important details. These pieces are then processed by existing 2D models without retraining them, making analysis more efficient and effective. This method works well across several different brain MRI datasets and adapts to data from other sources too.
Open → 2609.35405v1

Functional neural network models brain imaging patterns on curved matrix space

Geometric Feature Learning for Functional Data Valued on the Symmetric Positive Definite Manifold

Abstract: We here develop a functional neural network, termed MatFAE, for learning trajectories on the Riemannian manifold of symmetric positive definite (SPD) matrices. MatFAE features intrinsic layers that map manifold-valued functions to Euclidean vector-valued functions, followed by a functional layer that projects them into a finite-dimensional Euclidean space. Unlike most neural networks for discrete-time sequences, MatFAE treats each sequence as a continuous function and can therefore encode trajectory dynamics (e.g., first-order derivatives) in its latent representations. Additionally, the morphology of the functional weights in the functional layer offers interpretability by revealing the regions of the input functional data that contribute most to the latent representations. We justify the design principles and properties of each intrinsic layer and detail how matrix factorization is handled during backpropagation. We apply MatFAE to a range of fMRI datasets, demonstrating its ability to efficiently learn informative representations from high-dimensional SPD trajectories and its practical value for real-world neuroimaging analysis.

Thu 24 SeptMachine Learning
The gist
The paper develops a special kind of neural network called MatFAE that learns from data which changes smoothly over time and lives on a curved space of matrices called SPD matrices. Unlike usual methods that see data as discrete points, MatFAE treats data as continuous functions, capturing how they evolve and revealing important parts of the data. The authors show how this method can analyze complex brain imaging data to find useful representations for further study in neuroscience.
Open → 2609.30487v1

Mesh guided repairs improve broken vessel segmentations connectivity

Mind the Gap: Mesh-Guided Repair of Broken Vessels

Abstract: Vessel segmentation is commonly optimized as voxel-wise classification, but small local errors can strongly disrupt vascular connectivity while having little effect on overlap scores. This is particularly problematic for downstream analyses that rely on centerlines, branches, connected components, or graph structure. We propose a mesh-guided post-processing framework for repairing broken vessel segmentations produced by nnU-Net. For each predicted binary mask, a deformable template mesh is fitted to the mask surface in physical space and used as a case-specific geometric scaffold. The fitted mesh is not voxelized as the final segmentation; instead, it guides conservative reconnection of disconnected components by proposing or validating thin bridge candidates under foreground-growth constraints. We evaluated this approach in three vascular anatomies using AortaSeg24 and SEGA for the aorta, TopCoW for the Circle of Willis, and PARSE for the pulmonary arteries. Performance is measured using Dice, connected-component Dice (ccDice), and the Betti-0 number. Across these datasets, repair substantially improved connectivity while preserving overlap: Dice remained nearly unchanged, whereas ccDice increased from 0.596 to 0.992 for aorta, from 0.722 to 0.835 for TopCoW, and from 0.028 to 0.862 for PARSE. The FOMAML meta-initialization further accelerated the fitting per-case, supporting practical mesh-based repair of the vascular topology. These results suggest that explicit mesh representations can provide a useful geometric prior for correcting topological failures in otherwise accurate voxel segmentations.

Thu 24 SeptComputer Vision and Pattern Recognition
The gist
Vessel images used by doctors and scientists sometimes have tiny breaks that mess up important structures without showing big mistakes in usual checks. The authors created a way to fix these breaks by fitting a flexible 3D mesh over the vessel shapes, then using the mesh to carefully reconnect the broken parts. They tested this on different artery datasets and found it made the vessel maps much more connected without losing accuracy. Their method runs efficiently and helps keep the important shapes intact for further analysis.
Open → 2609.29779v1