arXiv subjects

Jens Kleesiek

Publications and source records attributed to Jens Kleesiek.

At least 19 recordsLinked to original sources

arXiv2026

Modular Expert Merging for Biomedical Retrieval

Adapting general-purpose LLMs into domain-specialized dense retrievers typically requires large-scale training on mixed-domain data. We show that merging independently trained domain-specialized experts consistently exceeds this approach across four decoder-only LLM families (0.6B-7B), four merging methods, and twelve medical and general retrieval tasks from MTEB, suggesting that parameter-space composition captures complementary domain strengths that large-scale mixed-domain training averages out. To further maximize expert quality, we introduce Synthesize-Train-Merge (STM), a modular framework that synthesizes hard negatives with a top-tier LLM and fine-tunes domain-specialized experts via LoRA before merging them, without continual pre-training. Synthesized hard negatives yield the largest gains for smaller models, and STM achieves strong performance on biomedical retrieval tasks while maintaining competitive general-domain results across all four backbone families.

cs.CL↗

arXiv2026

SCoPE-Reg: Efficient Rigid Ultrasound Slice-to-Volume Registration via State-Space Correlation and Closed-Form Pose Estimation

Ultrasound-guided interventions can require localization of an untracked 2D frame within a 3D anatomical reference. Rigid slice-to-volume registration (SVR) estimates this six-degree-of-freedom pose but remains challenging because of limited anatomical context, acoustic artifacts, and view-dependent appearance. Existing methods often use dense cross-attention, whose cost scales with the product of slice and volume token counts, or direct pose regression without explicit correspondence constraints. We introduce SCoPE-Reg, combining state-space slice--volume interaction, dense 3D coordinate prediction, and parameter-free weighted Kabsch estimation. On SVR tasks from CAMUS and $μ$-RegPro, SCoPE-Reg yields mean target registration errors of $0.73$ mm and $2.27$ mm against $1.24$ mm and $2.63$ mm for the state of the art (SOTA), reduces peak error on CAMUS by 56% below SOTA ($12.5\!\to\!5.5$ mm), and registers $100\%$ and $80\%$ of frames within $3$ mm. On CAMUS at $128^2$ it retains the lowest error at increasing pose-perturbation magnitude. It holds $6.49$ M parameters independent of resolution, sustaining $51$ FPS at $512^2$. SCoPE-Reg establishes a SOTA in rigid ultrasound SVR: by coupling correspondence-based accuracy with bounded worst-case error and resolution-independent cost, it becomes viable at native acquisition resolution during intervention, where prior methods trade accuracy, reliability, or frame rate against one another. Supplementary code provided and will be open-sourced upon acceptance.

eess.IV↗

arXiv2026

Case2Flow: Bridging Patient Cases and Guideline Flowcharts through Multimodal Retrieval

Medical guidelines encode rich, evidence-based decision logic, yet the specific decision artifact a clinician needs is hard to locate within a guideline, let alone across guidelines covering plausible diseases and treatments. While guideline passages have supported end-to-end question answering, flowcharts remain largely underused in decision support despite their ability to encode actionable clinical pathways. We therefore introduce Case2Flow, a task designed to retrieve the most relevant guideline flowchart for a given patient case from a collection of guideline documents. To support it, we construct FlowAtlas, a curated corpus of 202 flowcharts extracted from 2,080 medical guidelines, together with a pipeline that synthesises 1,911 aligned case-flowchart pairs. Our evaluation of multimodal retrieval methods reveals systematic failure modes, including overreliance on keywords and spurious token-patch matches induced by uninformative background regions in flowcharts. Motivated by this, we propose CRISP, a training-free scoring method that sharpens late-interaction retrieval by suppressing uninformative patches, discounting ambiguous token matches, and incorporating bidirectional query-image alignment. CRISP improves Recall@1 by up to 18.71 percentage points, while a blinded physician assessment on published case narratives provides preliminary feasibility evidence beyond synthetic queries.

cs.CL↗

arXiv2026

Good Enough? An Investigation on the Impact of Label Quality in Large-Scale Medical Datasets

Manually refining radiological segmentation masks is highly resource-intensive. To determine when this expert commitment is truly justified for the training of segmentation models, we investigate the relationship between label quality and model performance. Expanding beyond models trained directly for inference, we conduct the first study isolating the impact of label quality in pre-training datasets. While high-quality labels remain essential for models proceeding directly to deployment, we find no evidence that strict label quality is crucial for pre-training efficacy. These results question the necessity of exhaustive human-in-the-loop refinement for massive corpora intended for pretraining and suggest that expert effort is more effectively invested in well-curated downstream target datasets.

cs.CV↗

arXiv2026

Configurable Clinical Information Extraction with Agentic RAG: What Works, What Breaks, and Why

Patient contexts span hundreds of heterogeneous documents and thousands of structured data points, yet the document-level metadata that AI systems need for retrieval and triage is absent or incomplete. Standard retrieval-augmented generation fails on this data, mishandling temporal reasoning, cross-document dependencies, and missing metadata. We deploy ACIE (Agentic Clinical Information Extraction) at University Medicine Essen: an on-premise agentic RAG pipeline that reasons over complete patient contexts and grounds every answer in source passages for clinician verification. We quantify the metadata gap, trace the architectural decisions it shaped, and evaluate extraction alongside an independent retrospective lymphoma registry study, in which nuclear-medicine physicians verify every extracted value against its cited sources. Across 7,326 judgments, clinicians accepted 96.5\% of extractions, with per-type acceptance ranging from 80\% to 99\%.

cs.AI↗

arXiv2026

++nnU-Net: Scaling nnU-Net with Prefix-Based Data Augmentation

The nnU-Net has demonstrated continuous success in medical segmentation tasks, which heavily rely on the availability and diversity of annotated biomedical data. However, assembling medical imaging cohorts remains challenging due to numerous factors such as privacy regulations and annotation costs. As a result, data augmentation plays a crucial role in increasing data availability while maintaining anatomical feasibility. Hence, we propose the ++nnU-Net, a novel data augmentation module based on image registration that operates prior to preprocessing and training take place. Our framework was evaluated across five different 2D datasets. In this workflow, image data go through a two-stage registration process, generating new warped images. The transformations are then applied to the respective segmentation. In addition, the pipeline computes available disk space, generates supplementary binary synthetic masks and generates checkpoints. We demonstrate that the ++nnU-Net outperforms the nnU-Net baseline, yielding improvements in Dice Similarity Coefficient scores. In the most prominent cases, we observe performance gains of approximately 22\%. These findings highlight the effectiveness of registration-based data augmentation, particularly for 2D medical imaging datasets and suggest that the ++nnU-Net provides a practical and scalable approach for enhancing segmentation performance in data-limited settings. The source code for the ++nnU-Net is available at: https://github.com/sofia-adelie/plusplusnnunet.git

eess.IV↗

arXiv2026

Robustness of breast lesion segmentation under MRI undersampling improves with k-space-aware deep learning

Purpose: To assess whether breast lesion segmentation can be learned directly from acquired MRI k-space, and whether doing so improves robustness when data are accelerated or noisy. Materials and Methods: This retrospective study used public breast dynamic contrast-enhanced MRI (DCE-MRI) datasets with acquired and synthetic k-space, together with a within-dataset synthetic control. We compared four 3D U-Net variants: a hybrid k-space-to-image model, a native k-space model, and magnitude and complex image-space baselines. Models were evaluated under increasing undersampling and added complex Gaussian k-space noise. The primary outcome was patient-level Dice similarity coefficient under cross-validation, with the hybrid model prespecified as the main comparison against the magnitude image-space baseline. Results: At full sampling, the hybrid and image-space models performed similarly. As acceleration increased, the hybrid model retained substantially more segmentation accuracy and significantly outperformed the magnitude image-space baseline across moderate to high undersampling levels. The same pattern was observed when noise was added directly to k-space: the hybrid model degraded more slowly, whereas the image-space baseline failed under heavier noise. This advantage was reproduced in the within-dataset synthetic control. Feature analysis suggested that the k-space stage and image-space stage played complementary roles, with frequency-domain filtering concentrated before image-domain lesion localization. Conclusion: K-space-aware deep learning improves the robustness of breast lesion segmentation under MRI undersampling and k-space noise, while matching image-space methods at full sampling.

cs.CV↗

arXiv2026

Beyond Parameter Aggregation: Semantic Consensus for Federated Fine-Tuning of LLMs

Federated fine-tuning of large language models is commonly formulated as a parameter aggregation problem. However, even parameter-efficient methods require transmitting large collections of trainable weights, assume aligned architectures, and rely on white-box access to model parameters. As model sizes continue to grow and deployments become increasingly heterogeneous, these assumptions become progressively misaligned with practical constraints. We consider an alternative formulation in which collaboration is mediated through model behavior rather than parameters. Clients fine-tune local models on private data and exchange generated outputs on a shared, public prompt set. The server maps these outputs into a semantic representation space, forms a per-prompt semantic consensus, and returns pseudo-labels for further local fine-tuning. This formulation fundamentally changes the communication scaling of federated LLM fine-tuning. The amount of information exchanged depends only on the public prompt budget and the size of the communicated behaviors, independent of model size. As a consequence, the protocol naturally accommodates heterogeneous architectures and applies directly to open-ended text generation. We present a theoretical analysis and empirical results demonstrating that this approach can match strong federated fine-tuning baselines while substantially reducing communication by orders of magnitude (e.g., analytically by a factor of $1006$ for Llama3.1-405B), as well as reductions in runtime and energy consumption. These results suggest that, for generative foundation models, behavior-level consensus provides a more appropriate abstraction for federated adaptation than parameter aggregation.

cs.LG↗

arXiv2026

The autoPET3 Challenge: Automated Lesion Segmentation in Whole-Body PET/CT $\unicode{x2013}$ Multitracer Multicenter Generalization

We report the design and results of the third autoPET challenge (MICCAI 2024), which benchmarked automated lesion segmentation in whole-body PET/CT under a compositional generalization setting. Training data comprised 1,014 [18F]-FDG PET/CT studies from the University Hospital Tübingen and 597 [18F]/[68Ga]-PSMA PET/CT studies from the LMU University Hospital Munich, constituting the largest publicly available annotated PSMA PET/CT dataset to date. The held-out test set of 200 studies covered four tracer-center combinations, two of which represented unseen compositional pairings. A complementary data-centric award category isolated the contribution of data handling strategies by restricting participants to a fixed baseline model. Seventeen teams submitted 27 algorithms, predominantly nnU-Net-based 3D networks with PET/CT channel concatenation. The top-ranked algorithm achieved a mean DSC of 0.66, FNV of 3.18 mL, and FPV of 2.78 mL across all four test conditions, improving DSC by 8% and reducing the false-negative volume by 5 mL relative to the provided baseline. Ranking was stable across bootstrap resampling and alternative ranking schemes for the top tier. Beyond the benchmark, we provide an in-depth analysis of segmentation performance at the patient and lesion level. Three main conclusions can be drawn: (1) in-domain multitracer PET/CT segmentation is sufficient and probably approaching reader agreement; (2) compositional generalization to unseen tracer-center combinations remains an open problem mainly driven by systematic volume overestimation; (3) heterogeneity and case difficulty drive performance variation substantially more than the choice of algorithm among top-ranked teams.

cs.CV↗

arXiv2026

VS-DDPM: Efficient Low-Cost Diffusion Model for Medical Modality Translation

Diffusion models produce high-quality synthetic data but suffer from slow inference. We propose 3D Variable-Step Denoising Diffusion Probabilistic Model (VS-DDPM) a framework engineered to maintain generative quality while accelerating inference by several factors. We tested our approach on four tasks (missing MRI, tumor removal, MRI-to-sCT, and CBCT-to-sCT) within the BraTS2025 and SynthRAD2025 challenges. Designed for high efficiency under hardware and time constrains imposed by both challenges. VS-DDPM achieved state-of-the-art (SOTA) performance in missing MRI synthesis, yielding Dice scores of 0.80, 0.83, and 0.88 for the enhancing tumor, tumor core, and whole tumor regions, respectively, alongside a structural similarity index (SSIM) of 0.95. For MRI tumor removal, the model attained a root mean squared error (RMSE) of 0.053, a peak signal-to-noise ratio (PSNR) of 26.77, and an SSIM of 0.918. While the framework demonstrated competitive performance in MRI-to-sCT and CBCT-to-sCT tasks, it did not reach SOTA benchmarks, potentially due to sensitivities in data pre and post-processing pipelines or specific loss function configurations. These results demonstrate that VS-DDPM provides a robust and tunable solution for high-fidelity 3D medical image synthesis. The code is available in https://github.com/andre-fs-ferreira/SynthRAD_by_Faking_it.

cs.CV↗

arXiv2026

Generative Modeling of Complex-Valued Brain MRI Data

Objective. Standard Magnetic Resonance Imaging (MRI) reconstruction pipelines discard phase information captured during acquisition, despite evidence that it encodes tissue properties relevant to tumor diagnosis. Current machine learning approaches inherit this limitation by operating exclusively on reconstructed magnitude images. The aim of this study is to build a generative framework which is capable of jointly modeling magnitude and phase information of complex-valued MRI scans. Approach. The proposed generative framework combines a conditional variational autoencoder, which compresses complex-valued MRI scans into compact latent representations while preserving phase coherence, with a flow-matching-based generative model. Synthetic sample quality is assessed via a real-versus-synthetic classifier and by training downstream classifiers on synthetic data for abnormal tissue detection. Main results. The autoencoder preserves phase coherence above 0.997. Real-versus-synthetic classification yields low AUROC values between 0.50 and 0.66 across all acquisition sequences, indicating generated samples are nearly indistinguishable from real data. In downstream normal-versus-abnormal classification, classifiers trained entirely on synthetic data achieve an AUROC of 0.880, surpassing the real-data baseline of 0.842 on a publicly available dataset (fastMRI). This advantage persists on an independent external test set from a different institution with biopsy-confirmed labels. Significance. The proposed framework demonstrates the feasibility of jointly modeling magnitude and phase information for normal and abnormal complex-valued brain MRI data. Beyond synthetic data generation, it establishes a foundation for the usage of complete brain MRI information in future diagnostic applications and enables systematic investigation of how magnitude and phase jointly encode pathology-specific features.

eess.IV↗

arXiv2026

Learning to Look Closer: A New Instance-Wise Loss for Small Cerebral Lesion Segmentation

Traditional loss functions in medical image segmentation, such as Dice, often under-segment small lesions because their small relative volume contributes negligibly to the overall loss. To address this, instance-wise loss functions and metrics have been proposed to evaluate segmentation quality on a per-lesion basis. We introduce CC-DiceCE, a loss function based on the CC-Metrics framework, and compare it with the existing blob loss. Both are benchmarked against a DiceCE baseline within the nnU-Net framework, which provides a robust and standardized setup. We find that CC-DiceCE loss increases detection (recall) with minimal to no degradation in segmentation performance, though with dataset-dependent trade-offs in precision. Furthermore, our multi-dataset study shows that CC-DiceCE generally outperforms blob loss.

cs.CV↗

arXiv2026

Flatness is Necessary, Neural Collapse is Not: Rethinking Generalization via Grokking

Neural collapse, i.e., the emergence of highly symmetric, class-wise clustered representations, is frequently observed in deep networks and is often assumed to reflect or enable generalization. In parallel, flatness of the loss landscape has been theoretically and empirically linked to generalization. Yet, the causal role of either phenomenon remains unclear: Are they prerequisites for generalization, or merely by-products of training dynamics? We disentangle these questions using grokking, a training regime in which memorization precedes generalization, allowing us to temporally separate generalization from training dynamics and we find that while both neural collapse and relative flatness emerge near the onset of generalization, only flatness consistently predicts it. Models encouraged to collapse or prevented from collapsing generalize equally well, whereas models regularized away from flat solutions exhibit delayed generalization, resembling grokking, even in architectures and datasets where it does not typically occur. Furthermore, we show theoretically that neural collapse leads to relative flatness under classical assumptions, explaining their empirical co-occurrence. Our results support the view that relative flatness is a potentially necessary and more fundamental property for generalization, and demonstrate how grokking can serve as a powerful probe for isolating its geometric underpinnings.

cs.LG↗

arXiv2026

AIANO: Enhancing Information Retrieval with AI-Augmented Annotation

The rise of Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) has rapidly increased the need for high-quality, curated information retrieval datasets. These datasets, however, are currently created with off-the-shelf annotation tools that make the annotation process complex and inefficient. To streamline this process, we developed a specialized annotation tool - AIANO. By adopting an AI-augmented annotation workflow that tightly integrates human expertise with LLM assistance, AIANO enables annotators to leverage AI suggestions while retaining full control over annotation decisions. In a within-subject user study ($n = 15$), participants created question-answering datasets using both a baseline tool and AIANO. AIANO nearly doubled annotation speed compared to the baseline while being easier to use and improving retrieval accuracy. These results demonstrate that AIANO's AI-augmented approach accelerates and enhances dataset creation for information retrieval tasks, advancing annotation capabilities in retrieval-intensive domains.

cs.IR↗

arXiv2026

Region-Normalized DPO for Medical Image Segmentation under Noisy Judges

While dense pixel-wise annotations remain the gold standard for medical image segmentation, they are costly to obtain and limit scalability. In contrast, many deployed systems already produce inexpensive automatic quality-control (QC) signals like model agreement, uncertainty measures, or learned mask-quality scores which can be used for further model training without additional ground-truth annotation. However, these signals can be noisy and biased, making preference-based fine-tuning susceptible to harmful updates. We study Direct Preference Optimization (DPO) for segmentation from such noisy judges using proposals generated by a supervised base segmenter trained on a small labeled set. We find that outcomes depend strongly on how preference pairs are mined: selecting the judge's top-ranked proposal can improve peak performance when the judge is reliable, but can amplify harmful errors under weaker judges. We propose Region-Normalized DPO (RN-DPO), a segmentation-aware objective which normalizes preference updates by the size of the disagreement region between masks, reducing the leverage of harmful comparisons and improving optimization stability. Across two medical datasets and multiple regimes, RN-DPO improves sustained performance and stabilizes preference-based fine-tuning, outperforming standard DPO and strong baselines without requiring additional pixel annotations.

cs.CV↗

arXiv2026

Efficient Complex-Valued Vision Transformers for MRI Classification Directly from k-Space

Deep learning applications in Magnetic Resonance Imaging (MRI) predominantly operate on reconstructed magnitude images, a process that discards phase information and requires computationally expensive transforms. Standard neural network architectures rely on local operations (convolutions or grid-patches) that are ill-suited for the global, non-local nature of raw frequency-domain (k-Space) data. In this work, we propose a novel complex-valued Vision Transformer (kViT) designed to perform classification directly on k-Space data. To bridge the geometric disconnect between current architectures and MRI physics, we introduce a radial k-Space patching strategy that respects the spectral energy distribution of the frequency-domain. Extensive experiments on the fastMRI and in-house datasets demonstrate that our approach achieves classification performance competitive with state-of-the-art image-domain baselines (ResNet, EfficientNet, ViT). Crucially, kViT exhibits superior robustness to high acceleration factors and offers a paradigm shift in computational efficiency, reducing VRAM consumption during training by up to 68$\times$ compared to standard methods. This establishes a pathway for resource-efficient, direct-from-scanner AI analysis.

cs.CV↗

arXiv2025

Does Biomedical Training Lead to Better Medical Performance?

Large Language Models (LLMs) are expected to significantly contribute to patient care, diagnostics, and administrative processes. Emerging biomedical LLMs aim to address healthcare-specific challenges, including privacy demands and computational constraints. Assessing the models' suitability for this sensitive application area is of the utmost importance. However, biomedical training has not been systematically evaluated on medical tasks. This study investigates the effect of biomedical training in the context of six practical medical tasks evaluating $25$ models. In contrast to previous evaluations, our results reveal a performance decline in nine out of twelve biomedical models after fine-tuning, particularly on tasks involving hallucinations, ICD10 coding, and instruction adherence. General-domain models like Meta-Llama-3.1-70B-Instruct outperformed their biomedical counterparts, indicating a trade-off between domain-specific fine-tuning and general medical task performance. We open-source all evaluation scripts and datasets at https://github.com/TIO-IKIM/CLUE to support further research in this critical area.

cs.CL↗

arXiv2025

Whom to Trust? Adaptive Collaboration in Personalized Federated Learning

Data heterogeneity poses a fundamental challenge in federated learning (FL), especially when clients differ not only in distribution but also in the reliability of their predictions across individual examples. While personalized FL (PFL) aims to address this, we observe that many PFL methods fail to outperform two necessary baselines, local training and centralized training. This suggests that meaningful personalization only emerges in a narrow regime, where global models are insufficient, but collaboration across clients still holds value. Our empirical findings point to two key ingredients for success in this regime: adaptivity in collaboration and fine-grained trust, at the level of individual examples. We show that these properties can be achieved within federated semi-supervised learning, where clients exchange predictions over a shared unlabeled dataset. This enables each client to align with public consensus when it is helpful, and disregard it when it is not, without sharing model parameters or raw data. As a concrete realization of this idea, we develop FEDMOSAIC, a personalized co-training method where clients reweight their loss and their contribution to pseudo-labels based on per-example agreement and confidence. FEDMOSAIC outperforms strong FL and PFL baselines across a range of non-IID settings, and we prove convergence under standard smoothness, bounded-variance, and drift assumptions. In contrast to many of these baselines, it also outperforms local and centralized training. These results clarify when federated personalization can be effective, and how fine-grained, trust-aware collaboration enables it.

cs.LG↗