arXiv subjects

Morten Rieger Hannemose

Publications and source records attributed to Morten Rieger Hannemose.

15 recordsLinked to original sources

arXiv2026

Structural Limits of the Information-Theoretic Uncertainty Decomposition

Uncertainty estimation in machine learning typically decomposes uncertainty into aleatoric uncertainty (AU) and epistemic uncertainty (EU) using the standard information-theoretic framework. However, in practice, two critical issues arise: entanglement (AU and EU are highly correlated) and epistemic collapse (EU magnitude shrinks with increasing model capacity). We analyze this framework on a functional level and discover that significant portions of the assumed AU, EU range are infeasible in finite settings, and cannot be attained with any class probabilities. We characterize how this infeasible region scales with the number of classes and Monte Carlo samples $N$ (e.g., from ensembles with $N$ members), revealing it is bounded by $\text{AU} \leq \log(2)/N$. Crucially, the infeasible region's boundary helps explain epistemic collapse: when model confidence is high, $\text{AU} > \text{EU}$ is guaranteed by this fundamental structural limitation. Our findings show that increasing ensemble size mitigates epistemic collapse by reducing the infeasible area. Lastly, we caution against interpreting AU and EU as independent quantities in low AU regimes, since we show they are coupled when $\text{AU} \leq \log(2)/N$.

cs.CV↗

arXiv2026

Rethinking Uncertainty Quantification and Entanglement in Image Segmentation

Uncertainty quantification (UQ) is crucial in safety-critical applications such as medical image segmentation. Total uncertainty is typically decomposed into data-related aleatoric uncertainty (AU) and model-related epistemic uncertainty (EU). Many methods exist for modeling AU (such as Probabilistic UNet, Diffusion) and EU (such as ensembles, MC Dropout), but it is unclear how they interact when combined. Additionally, recent work has revealed substantial entanglement between AU and EU, undermining the interpretability and practical usefulness of the decomposition. We present a comprehensive empirical study covering a broad range of AU-EU model combinations, propose an entanglement proxy based on the relative performance of uncertainty measures, and evaluate model combinations across downstream uncertainty quantification tasks. Ensembles consistently show more favorable proxy values and superior performance. Softmax models usually beat other AU methods, except in calibration where the results are dataset-dependent. A softmax ensemble performs remarkably well on all tasks. Finally, we analyze potential sources of uncertainty entanglement and outline directions for mitigating this effect.

cs.CV↗

arXiv2026

MedDiME: Efficient Latent Diffusion with Adaptive Masking for Medical Counterfactual Generation

Medical counterfactual generation modifies images to change model predictions for interpretability. However, existing diffusion-based approaches are often prohibitively slow and memory-intensive, making them difficult to apply in high-resolution settings. Moreover, existing masking strategies are tightly coupled with pixel-space representations, making them incompatible with latent-space diffusion editing. To address these challenges, we propose MedDiME, a latent-space classifier-guided diffusion framework that reduces computational and memory overhead while introducing a latent-compatible, gradient-driven adaptive masking mechanism for spatially precise medical counterfactual generation. Extensive experiments demonstrate that MedDiME achieves high-quality counterfactual generation with significant efficiency gains compared to prior classifier-guided diffusion baselines, achieving up to 40 times faster inference and 13 times lower peak GPU memory usage.

cs.CV↗

arXiv2026

FiRe: Fixed-Noise Refinement for Visual Counterfactual Explanations

Visual counterfactual explanations aim to change classifier decisions through realistic and localized edits while preserving decision-irrelevant content. Existing DDPM-based methods typically perform classifier-guided editing along a long reverse denoising trajectory. The changing noise levels make semantic editability and spatial control difficult to balance, and the editable state is noisy, whereas the target classifier is trained on clean images. As a result, these methods require either costly recursive denoising or low-quality one-step estimates to obtain classifier-facing clean images. We propose FiRe, a Fixed-noise Refinement framework for visual counterfactual explanations. Rather than following a reverse denoising trajectory, FiRe maps the input to a fixed noise level and iteratively refines the noisy state at that level. To provide clean images for classifier guidance, FiRe first adapts Pixel Mean Flow to visual counterfactual explanation, enabling direct clean-image prediction from noisy states. To make fixed-noise refinement produce minimal and localized counterfactual edits, FiRe introduces three FiRe-specific controls: a dynamic dual-mask strategy, adaptive guidance, and early stopping, which determine where edits accumulate, which changes become visible, and when refinement stops. Experiments on five tasks across three datasets show that, compared with the strongest recent baseline, FiRe achieves about 3$\times$ faster online inference and 8$\times$ fewer FLOPs while obtaining comparable or state-of-the-art counterfactual quality.

cs.CV↗

arXiv2026

MMLANDMARKS: a Cross-View Instance-Level Benchmark for Geo-Spatial Understanding

Geo-spatial analysis of our world benefits from a multimodal approach, as every single geographic location can be described in numerous ways (images from various viewpoints, textual descriptions, geographic coordinates, etc.). Current benchmarks have limited coverage across modalities, leading to specialized models that perform well in their respective domains, but do not fully take advantage of other geo-spatial modalities. We introduce the Multi-Modal Landmark dataset (MMLandmarks), a benchmark composed of four modalities: 197k high-resolution aerial images, 329k ground-view images, textual information, and geographic coordinates for 18.557 distinct landmarks in the United States. The MMLandmarks dataset has a one-to-one landmark level correspondence across every modality, which enables training and benchmarking models for various geo-spatial tasks, including cross-view Ground-to-Satellite retrieval, ground and satellite geolocalization, Text-to-Image, and Text-to-GPS retrieval. We show that current specialized and off-the-shelf foundation models cannot be trivially used to solve this variety of geo-spatial tasks, illustrating a gap where multimodal datasets lead to broader geo-spatial understanding. We employ a simple CLIP-inspired baseline that reflects versatility and broad generalization when trained with MMLandmarks.

cs.CV↗

arXiv2026

MaskDiME: Adaptive Masked Diffusion for Precise and Efficient Visual Counterfactual Explanations

Visual counterfactual explanations aim to reveal the minimal semantic modifications that can alter a model's prediction, providing causal and interpretable insights into deep neural networks. However, existing diffusion-based counterfactual generation methods are often computationally expensive, slow to sample, and imprecise in localizing the modified regions. To address these limitations, we propose MaskDiME, a simple, fast, yet effective diffusion framework that unifies semantic consistency and spatial precision through localized sampling. Our approach adaptively focuses on decision-relevant regions to achieve localized and semantically consistent counterfactual generation while preserving high image fidelity. Our training-free framework, MaskDiME, performs inference over 30x faster than the baseline and achieves comparable or state-of-the-art performance across five benchmark datasets spanning diverse visual domains, establishing a practical and generalizable solution for efficient counterfactual explanation.

cs.CV↗

arXiv2026

Learning to Build Shapes by Extrusion

We introduce Text Encoded Extrusions (TEE), a text-based representation that expresses mesh construction as sequences of face extrusions rather than polygon lists, and a method for generating 3D meshes from TEE using a large language model (LLM). By learning extrusion sequences that assemble a mesh, similar to the way artists create meshes, our approach naturally supports arbitrary output face counts and produces manifold meshes by design, in contrast to recent mesh generative transformer based models. The learnt extrusion sequences can also be applied to existing meshes - enabling editing in addition to generation. To train our model, we decompose a library of quadrilateral meshes with non-self-intersecting face loops into constituent loops, which can be viewed as their building blocks, and finetune an LLM on the steps for reassembling the quadrilateral meshes by performing a sequence of extrusions. We demonstrate that our representation enables reconstruction, novel shape synthesis, and the addition of new features to existing meshes.

cs.GR↗

arXiv2026

Towards Agnostic and Holistic Universal Image Segmentation with Bit Diffusion

This paper introduces a diffusion-based framework for universal image segmentation, making agnostic segmentation possible without depending on mask-based frameworks and instead predicting the full segmentation in a holistic manner. We present several key adaptations to diffusion models, which are important in this discrete setting. Notably, we show that a location-aware palette with our 2D gray code ordering improves performance. Adding a final tanh activation function is crucial for discrete data. On optimizing diffusion parameters, the sigmoid loss weighting consistently outperforms alternatives, regardless of the prediction type used, and we settle on x-prediction. While our current model does not yet surpass leading mask-based architectures, it narrows the performance gap and introduces unique capabilities, such as principled ambiguity modeling, that these models lack. All models were trained from scratch, and we believe that combining our proposed improvements with large-scale pretraining or promptable conditioning could lead to competitive models.

cs.CV↗

arXiv2025

SA-UNetv2: Rethinking Spatial Attention U-Net for Retinal Vessel Segmentation

Retinal vessel segmentation is essential for early diagnosis of diseases such as diabetic retinopathy, hypertension, and neurodegenerative disorders. Although SA-UNet introduces spatial attention in the bottleneck, it underuses attention in skip connections and does not address the severe foreground-background imbalance. We propose SA-UNetv2, a lightweight model that injects cross-scale spatial attention into all skip connections to strengthen multi-scale feature fusion and adopts a weighted Binary Cross-Entropy (BCE) plus Matthews Correlation Coefficient (MCC) loss to improve robustness to class imbalance. On the public DRIVE and STARE datasets, SA-UNetv2 achieves state-of-the-art performance with only 1.2MB memory and 0.26M parameters (less than 50% of SA-UNet), and 1 second CPU inference on 592 x 592 x 3 images, demonstrating strong efficiency and deployability in resource-constrained, CPU-only settings.

cs.CV↗

arXiv2025

Diffusion Based Ambiguous Image Segmentation

Medical image segmentation often involves inherent uncertainty due to variations in expert annotations. Capturing this uncertainty is an important goal and previous works have used various generative image models for the purpose of representing the full distribution of plausible expert ground truths. In this work, we explore the design space of diffusion models for generative segmentation, investigating the impact of noise schedules, prediction types, and loss weightings. Notably, we find that making the noise schedule harder with input scaling significantly improves performance. We conclude that x- and v-prediction outperform epsilon-prediction, likely because the diffusion process is in the discrete segmentation domain. Many loss weightings achieve similar performance as long as they give enough weight to the end of the diffusion process. We base our experiments on the LIDC-IDRI lung lesion dataset and obtain state-of-the-art (SOTA) performance. Additionally, we introduce a randomly cropped variant of the LIDC-IDRI dataset that is better suited for uncertainty in image segmentation. Our model also achieves SOTA in this harder setting.

cs.CV↗

arXiv2025

Med-Art: Diffusion Transformer for 2D Medical Text-to-Image Generation

Text-to-image generative models have achieved remarkable breakthroughs in recent years. However, their application in medical image generation still faces significant challenges, including small dataset sizes, and scarcity of medical textual data. To address these challenges, we propose Med-Art, a framework specifically designed for medical image generation with limited data. Med-Art leverages vision-language models to generate visual descriptions of medical images which overcomes the scarcity of applicable medical textual data. Med-Art adapts a large-scale pre-trained text-to-image model, PixArt-$α$, based on the Diffusion Transformer (DiT), achieving high performance under limited data. Furthermore, we propose an innovative Hybrid-Level Diffusion Fine-tuning (HLDF) method, which enables pixel-level losses, effectively addressing issues such as overly saturated colors. We achieve state-of-the-art performance on two medical image datasets, measured by FID, KID, and downstream classification performance.

cs.CV↗

arXiv2025

Two Views Are Better than One: Monocular 3D Pose Estimation with Multiview Consistency

Deducing a 3D human pose from a single 2D image is inherently challenging because multiple 3D poses can correspond to the same 2D representation. 3D data can resolve this pose ambiguity, but it is expensive to record and requires an intricate setup that is often restricted to controlled lab environments. We propose a method that improves the performance of deep learning-based monocular 3D human pose estimation models by using multiview data only during training, but not during inference. We introduce a novel loss function, consistency loss, which operates on two synchronized views. This approach is simpler than previous models that require 3D ground truth or intrinsic and extrinsic camera parameters. Our consistency loss penalizes differences in two pose sequences after rigid alignment. We also demonstrate that our consistency loss substantially improves performance for fine-tuning without requiring 3D data. Furthermore, we show that using our consistency loss can yield state-of-the-art performance when training models from scratch in a semi-supervised manner. Our findings provide a simple way to capture new data, e.g in a new domain. This data can be added using off-the-shelf cameras with no calibration requirements. We make all our code and data publicly available.

cs.CV↗

arXiv2023

SportsPose -- A Dynamic 3D sports pose dataset

Accurate 3D human pose estimation is essential for sports analytics, coaching, and injury prevention. However, existing datasets for monocular pose estimation do not adequately capture the challenging and dynamic nature of sports movements. In response, we introduce SportsPose, a large-scale 3D human pose dataset consisting of highly dynamic sports movements. With more than 176,000 3D poses from 24 different subjects performing 5 different sports activities, SportsPose provides a diverse and comprehensive set of 3D poses that reflect the complex and dynamic nature of sports movements. Contrary to other markerless datasets we have quantitatively evaluated the precision of SportsPose by comparing our poses with a commercial marker-based system and achieve a mean error of 34.5 mm across all evaluation sequences. This is comparable to the error reported on the commonly used 3DPW dataset. We further introduce a new metric, local movement, which describes the movement of the wrist and ankle joints in relation to the body. With this, we show that SportsPose contains more movement than the Human3.6M and 3DPW datasets in these extremum joints, indicating that our movements are more dynamic. The dataset with accompanying code can be downloaded from our website. We hope that SportsPose will allow researchers and practitioners to develop and evaluate more effective models for the analysis of sports performance and injury prevention. With its realistic and diverse dataset, SportsPose provides a valuable resource for advancing the state-of-the-art in pose estimation in sports.

cs.CV↗

arXiv2022

Was that so hard? Estimating human classification difficulty

When doctors are trained to diagnose a specific disease, they learn faster when presented with cases in order of increasing difficulty. This creates the need for automatically estimating how difficult it is for doctors to classify a given case. In this paper, we introduce methods for estimating how hard it is for a doctor to diagnose a case represented by a medical image, both when ground truth difficulties are available for training, and when they are not. Our methods are based on embeddings obtained with deep metric learning. Additionally, we introduce a practical method for obtaining ground truth human difficulty for each image case in a dataset using self-assessed certainty. We apply our methods to two different medical datasets, achieving high Kendall rank correlation coefficients, showing that we outperform existing methods by a large margin on our problem and data.

cs.CV↗

arXiv2022

Multi-modal data generation with a deep metric variational autoencoder

We present a deep metric variational autoencoder for multi-modal data generation. The variational autoencoder employs triplet loss in the latent space, which allows for conditional data generation by sampling in the latent space within each class cluster. The approach is evaluated on a multi-modal dataset consisting of otoscopy images of the tympanic membrane with corresponding wideband tympanometry measurements. The modalities in this dataset are correlated, as they represent different aspects of the state of the middle ear, but they do not present a direct pixel-to-pixel correlation. The approach shows promising results for the conditional generation of pairs of images and tympanograms, and will allow for efficient data augmentation of data from multi-modal sources.

eess.IV↗