HAL: Hyper Article en Ligne
Not a member yet
3159010 research outputs found
Sort by
Beyond Log Scales: Toward Cognitively Informed Bar Charts for Orders of Magnitude Values
No full textInternational audienceIn this work, we challenge the dominant use of logarithmic scales to communicate values spanning multiple orders of magnitude—Orders of Magnitude Values (OMVs)—to the general public. Focusing on bar charts, we incorporate cognitive insights into visualization design to better align with how humans perceive OMVs. Studies in cognitive psychology suggest that, for large numerical ranges such as millions and billions, people do not think logarithmically. Instead, they perceive numbers in a piecewise linear manner, grouping values into scale words (e.g., millions) and applying linear reasoning within each group. We build upon a recently introduced piecewise linear scale, EplusM, and validate its use in bar charts, which we refer to as EplusM bar charts. We also introduce two novel variants of the EplusM bar chart informed by findings in numerical perception: Bricks, which builds on the concepts of round numbers and subitizing, and Multi-Magnitude, which leverages categorical perception of large numbers. In a crowdsourced experiment, we evaluate four bar chart designs: 1) Log, 2) EplusM, 3) Bricks, and 4) Multi-Magnitude, across value retrieval and quantitative comparison tasks. Our results show that EplusM bar charts are significantly preferred over logarithmic designs, increase user confidence, and reduce perceived mental demand, while maintaining task performance. These findings suggest that EplusM bar charts can serve as effective alternatives to logarithmic ones when visualizing OMVs for general audiences
Tuberculosis detection on chest X-rays using two-dimensional multiscale symbolic dynamic entropy
No full textAvailable online 14 July 2025International audienceSeveral radiological patterns associated with pulmonary tuberculosis (TB) have been identified on chest X-rays (CXR) used for screening purposes. As a result, several automatic computational tools have emerged for this purpose. We propose a new algorithm, two-dimensional multiscale symbolic dynamic entropy (MSDE), to develop a computational tool sensitive to these subtle patterns variations and noise robustness for evaluating CXR images from healthy and TB-diagnosed individuals. The one-dimensional SDE algorithm was previously shown to be more efficient in detecting amplitude variations and in computational calculations (compared to other entropy algorithms). Additionally, we also extracted first-order statistical parameters like standard deviation (SD), and mean of positive pixels (MPP), among others. These MSDE and first-order texture features were used to detect TB in each lung individually. The MSDE was validated using a synthetic dataset and optimized for the best set of parameters. We verified that, for both lungs, the MSDE values were significantly different between healthy and TB CXR images (), and the effect size was d 0.23. From the first-order parameters, only the mean, SD, entropy, and MPP were statistically different between both groups for the left lung (; d 0.22). For the right lung, all first-order features significantly differentiated TB patients (; d 0.28). Finally, we show that a multi-layer perceptron obtained 86.4 and 85.2% accuracy in detecting TB in the left and right lungs, respectively. The highest sensitivity values achieved in this study were 71.4% and 81.8% for the left and right lungs, respectively
Photometric Virtual Visual Servoing based on Gaussian Splatting
No full textInternational audienceWe present a novel approach that integrates pho-tometric Image-Based Virtual Visual Servoing (IBVVS) withGaussian Splatting (GS), a recent and efficient 3D representa-tion for photo-realistic view synthesis. In our framework, theservoing process is performed entirely in simulation using aGS model trained on a sparse set of images from a scenewheras unseen images serve as target views for photometricIBVVS. At each iteration a rendered image from the GS modelsimulate the camera’s current view, and the pixel-wise intensityerror between the rendered and target images is used tocompute control commands for camera pose optimization. Thisframework removes the need for externally acquired explicit 3Dgeometry or precomputed dense depth maps from traditionalsensors, since depth information is implicitly obtained from theGS representation and used directly in the control loop. Themethod enables virtual servoing toward novel views that werenot captured during training. Experimental results demonstrateaccurate and smooth convergence, highlighting the potentialof learned view synthesis for 3D camera tracking and visualservoing applications
RGB-D Fusion for Wide Field of View User Feedback in Teleoperation Context
No full textInternational audienceEffective teleoperation involves immersive and responsive visual feedback to support depth perception and spatial understanding to achieve precise control. Standard camera views naturally constrain the operator's field-of-view (FoV) of the remote scene, especially in cluttered or dynamic scenarios. We present a real-time RGB-D fusion system that expands the operator's FoV by employing immersive 3D reconstruction. Our system incorporates the Azure Kinect sensor into Unreal Engine using the ROS communication, rendering live sensor information onto a spherical mesh. This allows for smooth, wide-FoV rendering of the scene with greater peripheral context and depth continuity. In contrast to planar or depth-free systems, the proposed method is enhanced by live depth retranscription for more interactive teleoperation, leading to better scene understanding. This architecture lays the basis for flexible, high-fidelity remote interaction for robotics applications. All our developments and implementations are publicly available at: https://github.com/isri-aist/RGB-D_Fusion
Unsupervised anomaly detection in brain FDG PET with deep generative models: An experimental analysis of model variability and mitigation strategies
No full textInternational audienceUnsupervised anomaly detection allows identifying anomalies from unlabeled data, making it useful for neuroimaging analysis and computer-aided diagnosis. Given an individual's scan, we use a generative model to construct a subject-specific image of healthy appearance and compare both images to spot anomalies. Designing anomaly maps in such way has drawbacks as the reconstructions are imperfect and some variability is not taken into account. We study model variability arising from using different random seeds during training and explore solutions to mitigate the effect of unwanted reconstruction errors and variability. Our experiments on 3D brain FDG PET scans from ADNI suggest that variance between models can be reduced by aggregating their reconstructions in a Z-score based anomaly map, or normalizing the anomaly map with a healthy validation set
Molecular dynamics study on the coupled effects of size and pre-existing oxide layer on the compressive mechanical properties of copper nanowires
No full textInternational audienceCopper nanowires generally exhibit a native oxide shell layer, which can significantly impact their performance and reliability, especially in nanoelectronics applications. Using molecular dynamics simulations with the variable charge ReaxFF potential, we systematically examine the effects of preexisting oxide layers on the mechanical properties and deformation mechanisms of [001]-oriented Cu nanowires with varying diameters at room temperature. Our findings reveal a size-dependent influence of the native oxide layer on the mechanical behavior. Specifically, the formation of an oxide shell (CuxOy) around the Cu core reduces the activation barrier for defect nucleation, reducing yield properties and, thereby, weakening the nanowires. This effect is more pronounced in smaller samples due to the intensified interaction between the metallic core and the oxide shell. Additionally, while the strength, elastic modulus, and yield stress increase with the diameter of pristine and oxidized specimens, pristine nanowires consistently exhibit superior mechanical properties when compared to their oxidized counterparts. The degradation in mechanical performance primarily stems from the early onset of plasticity initiated at the oxidized surface. These findings emphasize the detrimental impact of native oxide layers on the mechanical behavior of Cu nanowires and highlight the critical role played by size upon the mechanical properties of nano-oxidized metal samples. This work provides valuable insights into tailoring the mechanical properties of Cu nanowires, contributing to the optimization of their performance in both nanoelectronics and mechanical applications.</div
La condamnation de Javier Bolsonaro et la capacité de résistance de l’État de droit au Brésil
No full textInternational audienc
Mission accomplished? A post-assessment of EU ETS impact on power sector emissions reduction
No full textInternational audienceThe debate on the capacity of the European Union Emissions Trading System (EU ETS) to effectively induce CO emissions reduction is still ongoing. This is particularly noteworthy in the case of the power sector, where numerous decarbonization policies overlap. This paper contributes to this discussion by leveraging a methodological approach that circumvents the challenges of constructing credible counterfactuals for causal inference and allows for disentangling the impact of the EU ETS from other measures on the power sector’s abatement efforts, alongside influencing factors such as weather. Specifically, we employ a Bayesian structural time series (BSTS) model, conceptually related to synthetic control techniques, to assess the effectiveness of the three completed phases of the EU ETS (2005-2020) in reducing CO emissions in the power sector across 24 Member States. We analyze the policy implementation effect over the course of each phase by comparing actual power sector emissions with counterfactual estimates derived from contemporaneous predictors related to such emissions. The results indicate a statistically significant emissions reduction in the second and third phases, with no significant reduction in the first phase. The power sector’s centrality to the EU ETS, and its critical role in our economies emphasize the importance of our findings in evaluating emissions reduction objectives