Search

Background: Deep learning (DL)-based organ segmentation is increasingly used in radiotherapy, yet voxel-wise DL uncertainty maps are rarely presented to clinicians. Purpose: This study assessed how DL-generated uncertainty maps impact radiation oncologists during manual correction of prostate radiotherapy DL segmentations. Methods: Two nnUNet models were trained by 10-fold cross-validation on 434 MRI-only prostate cancer cases to segment the prostate and rectum. Each model was evaluated on 35 independent cases. Voxel-wise uncertainty was calculated using the SoftMax standard deviation (n=10) and visualized as a color-coded map. Four oncologists performed segmentation in two steps: Step 1: Rated segmentation quality and confidence using Likert scales and edited DL segmentations without uncertainty maps. Step 2 ($\geq 4$ weeks later): Repeated step 1, but with uncertainty maps available. Segmentation time was recorded for both steps, and oncologists provided qualitative free-text feedback. Histogram analysis compared voxel edits across uncertainty levels. Results: DL segmentations showed high agreement with oncologist edits. Quality ratings varied: rectum segmentation ratings slightly decreased overall in step 2, while prostate ratings differed among oncologists. Confidence ratings also varied. Three oncologists reduced segmentation time with uncertainty maps, saving 1-2 minutes per case. Histogram analysis showed 50% fewer edits for step 2 in low-uncertainty areas. Conclusions: Presenting DL segmentation uncertainty information to radiation oncologists influences their decision-making, quality perception, and confidence in the DL segmentations. Low-uncertainty regions were edited less frequently, indicating increased trust in DL predictions. Uncertainty maps improve efficiency by reducing segmentation time and can be a valuable clinical tool, enhancing radiotherapy planning efficiency.

Textual Concept Bottleneck Models (TBMs) are interpretable-by-design models for text classification that predict a set of salient concepts before making the final prediction. This paper proposes Complete Textual Concept Bottleneck Model (CT-CBM),a novel TCBM generator building concept labels in a fully unsupervised manner using a small language model, eliminating both the need for predefined human labeled concepts and LLM annotations. CT-CBM iteratively targets and adds important concepts in the bottleneck layer to create a complete concept basis and addresses downstream classification leakage through a parallel residual connection. CT-CBM achieves good results against competitors, offering a promising solution to enhance interpretability of NLP classifiers without sacrificing performance.

Flexible fibers at the microscopic scale, such as flagella and cilia, play essential roles in biological and synthetic systems. The dynamics of these slender filaments in viscous flows involve intricate interactions between their mechanical properties and hydrodynamic drag. In this paper, considering a 1D, planar, inextensible Euler-Bernoulli rod in a viscous fluid modeled by Resistive Force Theory, we establish the existence and uniqueness of solutions for the $N$-link model, a mechanical model, designed to approximate the continuous filament with rigid segments. Then, we prove the convergence of the $N$-link model's solutions towards the solutions to classical elastohydrodynamics equations of a flexible slender rod. This provides an existence result for the limit model, comparable to those by Mori and Ohm [Nonlinearity, 2023], in a different functional context and with different methods. Due to its mechanical foundation, the discrete system satisfies an energy dissipation law, which serves as one of the main ingredients in our proofs. Our results provide mathematical validation for the discretization strategy that consists in approximating a continuous filament by the mechanical $N$-link model, which does not correspond to a classical approximation of the underlying PDE.

Chronic kidney disease (CKD) is a major global health issue, affecting over 10% of the population and causing significant mortality. While kidney biopsy remains the gold standard for CKD diagnosis and treatment, the lack of comprehensive benchmarks for kidney pathology segmentation hinders progress in the field. To address this, we organized the Kidney Pathology Image Segmentation (KPIs) Challenge, introducing a dataset that incorporates preclinical rodent models of CKD with over 10,000 annotated glomeruli from 60+ Periodic Acid Schiff (PAS)-stained whole slide images. The challenge includes two tasks, patch-level segmentation and whole slide image segmentation and detection, evaluated using the Dice Similarity Coefficient (DSC) and F1-score. By encouraging innovative segmentation methods that adapt to diverse CKD models and tissue conditions, the KPIs Challenge aims to advance kidney pathology analysis, establish new benchmarks, and enable precise, large-scale quantification for disease research and diagnosis.

This paper presents a novel approach to dual-frequency comb generation utilizing a single fiber Fabry-Perot resonator, advancing the implementation of these sources in fiber-based systems. Dual-comb applications such as spectroscopy, ranging, and imaging, known for their high-resolution and rapid data acquisition capabilities, benefit significantly from the stability and coherence of optical frequency comb sources. Our method leverages the birefringent property of the resonator induced by the optical fiber to generate two orthogonally polarized optical frequency combs in a monolitic resonator. This approach allows for the generation of two different frequency combs with slightly different repetition rates, exhibiting excellent mutual coherence, making it highly relevant for dual-comb applications. The 40 nm bandwidth generated combs are induced by switching-waves in a normal dispersion fiber Fabry-Perot resonator. These comb types have the advantage of being easily generated by a pulse pumping scheme, which is employed in this study. Finally, the potential of the source is demonstrated by a proof-of-concept spectroscopy measurement.

Radiotherapy treatment for prostate cancer relies on computed tomography (CT) and/or magnetic resonance imaging (MRI) for segmentation of target volumes and organs at risk (OARs). Manual segmentation of these volumes is regarded as the gold standard for ground truth in machine learning applications but to acquire such data is tedious and time-consuming. A publicly available clinical dataset is presented, comprising MRI- and synthetic CT (sCT) images, target and OARs segmentations, and radiotherapy dose distributions for 432 prostate cancer patients treated with MRI-guided radiotherapy. An extended dataset with 35 patients is also included, with the addition of deep learning (DL)-generated segmentations, DL segmentation uncertainty maps, and DL segmentations manually adjusted by four radiation oncologists. The publication of these resources aims to aid research within the fields of automated radiotherapy treatment planning, segmentation, inter-observer analyses, and DL model uncertainty investigation. The dataset is hosted on the AIDA Data Hub and offers a free-to-use resource for the scientific community, valuable for the advancement of medical imaging and prostate cancer radiotherapy research., Comment: 4 figures

Memory retention mechanisms play a central role in determining the efficiency of computational architectures designed for processing extended sequences. Conventional methods for token management often impose fixed retention thresholds or rely on uniform attention weight distributions, leading to inefficient memory utilization and premature information loss in extended sequence modeling. Structured Token Retention (STR) introduces a probabilistic selection framework that dynamically adjusts token persistence based on contextual significance, ensuring that computational resources are allocated to semantically relevant elements. Computational Memory Paths (CMP) extend this framework through hierarchical memory allocation, refining retention efficiency through structured reallocation of token embeddings. Comparative assessments against baseline models demonstrate that STR and CMP improve token survival rates across long input sequences while reducing cumulative error propagation across processing layers. Experimental results further indicate reductions in computational overhead, improving inference speed without degrading contextual coherence. Token distribution analyses reveal that structured memory allocation prevents excessive redundancy in attention weight calculations, optimizing information retrieval efficiency in large-scale generative architectures. The integration of STR and CMP into an open-source model illustrates the adaptability of structured memory retention methodologies, highlighting their applicability in generative text processing, long-context comprehension, and scalable sequence modeling.

We report the generation of a stable, broadband frequency comb, covering more than 10 THz, using a normal dispersion fiber Fabry-Perot resonator with a high quality factor of 69 millions. This platform ensures robust and easy integration into photonic devices via FC/PC connectors, and feature quality factors comparable to those of microresonators. We demonstrate a passive mode-locking phenomenon induced by the coherent interaction of the Kerr effect and Brillouin scattering, which generates a frequency comb with a repetition rate exceeding the free spectral range of the cavity. This parametric process modulates the continuous wave (CW) pump and can then be transformed into a train of almost square-wave pulses thanks to the generation of switching waves. Our results are supported by advanced numerical simulations, and theoretical derivations that include the Brillouin effect in the Fabry-Perot configuration. The very high stable feature of this optical frequency comb lying in the GHz range is critical to several applications ranging from telecommunication, spectroscopy and advanced microwave generation.

Accurate depth information is essential for many computer vision applications. Yet, no available dataset recording method allows for fully dense accurate depth estimation in a large scale dynamic environment. In this paper, we introduce DOC-Depth, a novel, efficient and easy-to-deploy approach for dense depth generation from any LiDAR sensor. After reconstructing consistent dense 3D environment using LiDAR odometry, we address dynamic objects occlusions automatically thanks to DOC, our state-of-the art dynamic object classification method. Additionally, DOC-Depth is fast and scalable, allowing for the creation of unbounded datasets in terms of size and time. We demonstrate the effectiveness of our approach on the KITTI dataset, improving its density from 16.1% to 71.2% and release this new fully dense depth annotation, to facilitate future research in the domain. We also showcase results using various LiDAR sensors and in multiple environments. All software components are publicly available for the research community., Comment: Preprint. Code and dataset available on the project page : https://simondemoreau.github.io/DOC-Depth/

Accurate sensor calibration is crucial for autonomous systems, yet its uncertainty quantification remains underexplored. We present the first approach to integrate uncertainty awareness into online extrinsic calibration, combining Monte Carlo Dropout with Conformal Prediction to generate prediction intervals with a guaranteed level of coverage. Our method proposes a framework to enhance existing calibration models with uncertainty quantification, compatible with various network architectures. Validated on KITTI (RGB Camera-LiDAR) and DSEC (Event Camera-LiDAR) datasets, we demonstrate effectiveness across different visual sensor types, measuring performance with adapted metrics to evaluate the efficiency and reliability of the intervals. By providing calibration parameters with quantifiable confidence measures, we offer insights into the reliability of calibration estimates, which can greatly improve the robustness of sensor fusion in dynamic environments and usefully serve the Computer Vision community., Comment: Accepted for publication at WACV 2025

This paper is devoted to the class of paraconvex functions and presents some of its fundamental properties, characterization, and examples that can be used for their recognition and optimization. Next, the convergence analysis of the projected subgradient methods with several step-sizes (i.e., constant, nonsummable, square-summable but not summable, geometrically decaying, and Scaled Polyak's step-sizes) to global minima for this class of functions is studied. In particular, the convergence rate of the proposed methods is investigated under paraconvexity and the H\"{o}lderian error bound condition, where the latter is an extension of the classical error bound condition. The preliminary numerical experiments on several robust low-rank matrix recovery problems (i.e., robust matrix completion, image inpainting, robust nonnegative matrix factorization, robust matrix compression) indicate promising behavior for these projected subgradient methods, validating our theoretical foundations., Comment: 38 pages, 62 figures

Block-Term Tensor Regression (BTTR) has proven to be a powerful tool for modeling complex, high-dimensional data by leveraging multilinear relationships, making it particularly well-suited for applications in healthcare and neuroscience. However, traditional implementations of BTTR rely on centralized datasets, which pose significant privacy risks and hinder collaboration across institutions. To address these challenges, we introduce Federated Block-Term Tensor Regression (FBTTR), an extension of BTTR designed for federated learning scenarios. FBTTR enables decentralized data analysis, allowing institutions to collaboratively build predictive models while preserving data privacy and complying with regulations. FBTTR represents a major step forward in applying tensor regression to federated learning environments. Its performance is evaluated in two case studies: finger movement decoding from Electrocorticography (ECoG) signals and heart disease prediction. In the first case study, using the BCI Competition IV dataset, FBTTR outperforms non-multilinear models, demonstrating superior accuracy in decoding finger movements. For the dataset, for subject 3, the thumb obtained a performance of 0.76 $\pm$ .05 compared to 0.71 $\pm$ 0.05 for centralised BTTR. In the second case study, FBTTR is applied to predict heart disease using real-world clinical datasets, outperforming both standard federated learning approaches and centralized BTTR models. In the Fed-Heart-Disease Dataset, an AUC-ROC was obtained of 0.872 $\pm$ 0.02 and an accuracy of 0.772 $\pm$ 0.02 compared to 0.812 $\pm$ 0.003 and 0.753 $\pm$ 0.007 for the centralized model.

We demonstrate that including the second viscosity of an electron gas in the hydrodynamic model allows for highly accurate modeling of the optical response of heavily doped semiconductors. In our setup, which improves resonance visibility compared to previous approaches, plasmon resonances become more distinct, allowing for detailed analysis of the underlying physics. With advanced fitting techniques based on a physics-informed cost function and a tailored optimization algorithm, we obtain close agreement between simulations and experimental data across different sample thicknesses. This enhanced resonance visibility, combined with our integrated approach, shows that key parameters such as doping level and effective electron mass can be retrieved from a single optical measurement. The spatial dispersion taken into account in the hydrodynamic framework is essential for accurately describing the optical response of plasmonic materials in this frequency range and is likely to become a standard modeling approach.

Selecting an appropriate prior to compensate for information loss due to the measurement operator is a fundamental challenge in imaging inverse problems. Implicit priors based on denoising neural networks have become central to widely-used frameworks such as Plug-and-Play (PnP) algorithms. In this work, we introduce Fixed-points of Restoration (FiRe) priors as a new framework for expanding the notion of priors in PnP to general restoration models beyond traditional denoising models. The key insight behind FiRe is that natural images emerge as fixed points of the composition of a degradation operator with the corresponding restoration model. This enables us to derive an explicit formula for our implicit prior by quantifying invariance of images under this composite operation. Adopting this fixed-point perspective, we show how various restoration networks can effectively serve as priors for solving inverse problems. The FiRe framework further enables ensemble-like combinations of multiple restoration models as well as acquisition-informed restoration networks, all within a unified optimization approach. Experimental results validate the effectiveness of FiRe across various inverse problems, establishing a new paradigm for incorporating pretrained restoration models into PnP-like algorithms.

Scientists and engineers use simulators to model empirically observed phenomena. However, tuning the parameters of a simulator to ensure its outputs match observed data presents a significant challenge. Simulation-based inference (SBI) addresses this by enabling Bayesian inference for simulators, identifying parameters that match observed data and align with prior knowledge. Unlike traditional Bayesian inference, SBI only needs access to simulations from the model and does not require evaluations of the likelihood-function. In addition, SBI algorithms do not require gradients through the simulator, allow for massive parallelization of simulations, and can perform inference for different observations without further simulations or training, thereby amortizing inference. Over the past years, we have developed, maintained, and extended $\texttt{sbi}$, a PyTorch-based package that implements Bayesian SBI algorithms based on neural networks. The $\texttt{sbi}$ toolkit implements a wide range of inference methods, neural network architectures, sampling methods, and diagnostic tools. In addition, it provides well-tested default settings but also offers flexibility to fully customize every step of the simulation-based inference workflow. Taken together, the $\texttt{sbi}$ toolkit enables scientists and engineers to apply state-of-the-art SBI methods to black-box simulators, opening up new possibilities for aligning simulations with empirically observed data.

In this work we study the behavior of the forward-backward (FB) algorithm when the proximity operator is replaced by a sub-iterative procedure to approximate a Gaussian denoiser, in a Plug-and-Play (PnP) fashion. In particular, we consider both analysis and synthesis Gaussian denoisers within a dictionary framework, obtained by unrolling dual-FB iterations or FB iterations, respectively. We analyze the associated minimization problems as well as the asymptotic behavior of the resulting FB-PnP iterations. In particular, we show that the synthesis Gaussian denoising problem can be viewed as a proximity operator. For each case, analysis and synthesis, we show that the FB-PnP algorithms solve the same problem whether we use only one or an infinite number of sub-iteration to solve the denoising problem at each iteration. To this aim, we show that each "one sub-iteration" strategy within the FB-PnP can be interpreted as a primal-dual algorithm when a warm-restart strategy is used. We further present similar results when using a Moreau-Yosida smoothing of the global problem, for an arbitrary number of sub-iterations. Finally, we provide numerical simulations to illustrate our theoretical results. In particular we first consider a toy compressive sensing example, as well as an image restoration problem in a deep dictionary framework.

MXenes stand out from other 2D materials because they combine very good electrical conductivity with hydrophilicity, allowing cost-effective processing as thin films. Therefore, there is a high fundamental interest in unraveling the electronic transport mechanisms at stake in multilayers of the most conducting MXene, Ti$_3$C$_2$T$_x$. Although weak localization (WL) has been proposed as the dominating low-temperature (LT) transport mechanism in Ti$_3$C$_2$T$_x$ thin films, there have been few attempts to model it quantitatively. In this paper, the role of important structural parameters -- thickness, interflake coupling, defects -- on the dimensionality of the LT transport mechanisms in spin-coated Ti$_3$C$_2$T$_x$ thin films is investigated through LT and magnetic field dependent resistivity measurements. A dimensional crossover from 2D to 3D WL is clearly evidenced when the film thickness exceeds the dephasing length $l_\phi$, which is in the $50-100\,\mathrm{nm}$ range. 2D WL can be restored by weakening the coupling between adjacent flakes, the intrinsic thickness of which is lower than $l_\phi$, hence acting as parallel 2D conductors. Alternatively, $l_\phi$ can be reduced down to the $10\,\mathrm{nm}$ range by defects. Our results clearly emphasize the ability of WL quantitative study to give deep insights in the physics of electron transport in MXene thin films.

Traditionally, 3d indoor datasets have generally prioritized scale over ground-truth accuracy in order to obtain improved generalization. However, using these datasets to evaluate dense geometry tasks, such as depth rendering, can be problematic as the meshes of the dataset are often incomplete and may produce wrong ground truth to evaluate the details. In this paper, we propose SCRREAM, a dataset annotation framework that allows annotation of fully dense meshes of objects in the scene and registers camera poses on the real image sequence, which can produce accurate ground truth for both sparse 3D as well as dense 3D tasks. We show the details of the dataset annotation pipeline and showcase four possible variants of datasets that can be obtained from our framework with example scenes, such as indoor reconstruction and SLAM, scene editing & object removal, human reconstruction and 6d pose estimation. Recent pipelines for indoor reconstruction and SLAM serve as new benchmarks. In contrast to previous indoor dataset, our design allows to evaluate dense geometry tasks on eleven sample scenes against accurately rendered ground truth depth maps.

The present study investigates the behaviour of the far-field sound radiated by low Mach number tip clearance flow induced by placing a stationary cambered airfoil adjacent to a stationary wall. The tip clearance heights ranged from 14% to 30% of the incoming, undisturbed boundary layer thickness and the clearance heights based Reynolds numbers were between 2,600 and 16,000. The far-field sound measured using a microphone array was beamformed to reveal the dominant noise sources and how they behave when the flow Mach number, angle of attack and the clearance height were varied. The near-field behaviour was also examined through PIV measurements and surface pressure fluctuation measurements on the tip. The results show that the mid-to-high frequency noise generated by tip clearances is dominated by the leakage flow in the mid-chord and leading-edge regions, while a distinct low-frequency noise source with a different scaling behaviour exists close to the trailing-edge of the tip clearance. The origin of this low-frequency noise source is believed to be the tip separation vortex that resides close to the trailing-edge and induces significant turbulence levels in the region. The strength of this noise source decreases with clearance height which is consistent with a reduction in turbulence levels associated with the separation vortex. The magnitude of the mid-frequency clearance noise which scales with the sixth power of the Mach number, decreases with tip clearance height due to a reduction in the fluctuating pressure on the airfoil tip surface. The time-scale of this sound was independent of the flow velocity, implying that the source is non-compact. Smaller tip clearances were also found to generate louder high-frequency noise due to intense turbulence and pressure fluctuation levels concentrated near the leading-edge of the clearance., Comment: Submitted to Applied Acoustics Journal