HAL Portal UPHF (Université Polytechnique Hauts-de-France)
Not a member yet
28473 research outputs found
Sort by
Impact of video resolution on center of mass estimation: A cross-models analysis (preliminary study)
No full textInternational audienceRecent advances in markerless motion capture (MC) have made biomechanical analysis more accessible. However, video parameters might affect pose estimation accuracy. This study investigates how video quality (540p and 1080p at 16:9) influences 3D center of mass (CoM) accuracy across different triangulated 2D pose estimation algorithms (RTMpose, OpenPose, OpenPose with marker augmentation and Theia3D). A healthy participant (age: 52) performed 60 statics standings trials. CoM estimations from each algorithm were compared against Theia3D. Results indicate that video quality significantly affects estimation accuracy, with varying impacts across algorithms. Our findings suggest considering video settings in markerless MC.</div
Economie solidaire politique versus économie sociale d'entreprise. Deux conceptions antagonistes ?
No full textInternational audienc
Novel Diffuse Midline Glioma‐on‐Chip Recapitulating Tumor Biophysical Microenvironment to Assess the Heterogeneity of Response to Therapies
No full textInternational audienceDespite significant efforts, Diffuse Midline Gliomas (DMGs) remain incurable. Although promising results are obtained in preclinical studies, most approaches have failed to improve survival in these young patients. In this context, by integrating elements of the biophysical tumor microenvironment known to regulate the response to therapies, a new preclinical tool is developed to better evaluate the efficacy of antitumoral strategies. For this purpose, a novel DMG-on-Chip (DoC) is engineered composed of a 3D dense tumor disk embedded in an extracellular matrix, which is accessible for real-time monitoring using wide-field phase contrast or confocal fluorescence microscopy. By driving the oxygen supply within the chip solely in a radial manner, a hypoxia gradient is established that can be associated with changes in DMG cell phenotype, proliferation, and rewiring of metabolic and stress transcriptomic signatures. Finally, DoC is used to analyze the spatial heterogeneity of the response of DMG cell lines and patient-derived 3D cultures to treatments through cell segmentation. Altogether, these interdisciplinary characterizations validate the new tool as an interesting healthcare solution to understand how cell responses are modulated by biophysical or biochemical cues in the tumor microenvironment
Climat : au train où vont les choses...: Urgences environnementales et transports ferroviaires
No full textInternational audienc
Proposed Experiments for Detecting Contextual Hidden Variables
No full textInternational audienceWe propose two quantum experiments – modified Bell tests – that could detect contextual hidden variables underlying quantum mechanics. The experiments are inspired by hydrodynamic pilot-wave systems that mimic a wide range of quantum effects and exhibit a classical analog of contextuality. To justify the experiments, we show that contextual hidden variables are inevitable and ‘physics as usual’ if a unification between quantum mechanics and general relativity is possible. Accordingly, contextual theories can bypass Bell’s theorem in a way that is both local and non-conspiratorial. We end with a note on the relevance of exploratory experiments in the foundations of quantum physics
New definition of Grades of Operational and Tactical Automation in rail domain
No full textInternational audienceThe present paper puts forward a proposal for an extension to the conventional Grades of Automation (GoA) with a purpose of facilitating both dedicated operational and tactical assistance. The contribution aims to reintegrate the driver into the train driving activity with the new Grades of Operational and Tactical Automation (GOTA). The proposal of GOTA is intended to provide train drivers with the option of varying the level of automation over time. Furthermore, they will facilitate the breakdown of the assistance level into tactical and operational tasks. This breakdown will be according to the specific characteristics of the driver and a variety of criteria, taking into account human involvement and driving skills. The GOTA Selector which helps the driver to be optimally engaged while driving by adapting the GOTA to the specific needs is then planned
«Erreurs» de traduction, « intraduisibiles », décentrement de l'écriture et littérature-monde : Ahmadou Kourouma, Chinua Achebe, Wole Soyinka.
No full textMany studies point out “errors” in the translations of African literature, particularly due to insufficient knowledge of the linguistic background as well as the geohistorical and cultural contexts. These errors include unjustified semantic modifications, approximate translations, and mistranslations. However, this kind of “error” is not specific to the context of African literature. The translation of any literary work requires consideration of these different dimensions, whether African or not. Most of these studies assume that African literature is translatable in the traditional sense of the term. However, as Barbara Cassin argues, we will adopt the perspective that all literature is by definition “untranslatable.” The original is never completely translatable because of the change of language inherent to translation. Even so, this does not mean that the translation is necessarily inferior to the original. On the contrary, as Borges demonstrates in his Homeric versions or, more recently, in the context of Gérard Genette’s analyses, the work must be considered not only in its original version, but also through its various translations, which represent so many possible reinterpretations of it.In such a perspective, it is no longer a question of achieving the perfect translation, deemed impossible. Even if no “error” is committed in the traditional sense of the term, which can be assumed in the case of self-translation in particular, the fact remains that the best possible translation will always be different from the original. This is precisely what makes a translation interesting: it reveals that which can only appear through the translation of a work and which the original can only potentially contain. In addition, this notion of “error” in the traditional sense does not arise when these different translations are included in Goethe’s “Weltliteratur.” Similarly, the concept of “error” is also absent in the framework of “littérature-monde,” defined by Alain Mabanckou as encompassing marginalized literatures characterized by linguistic hybridity. It is in this light that the question of translation “errors” will be re-examined, through problematization rather than a definition of the “errors” committed.We will essentially rely on the English translations of two French-language works by the Ivorian author Ahmadou Kourouma, as well as the French translations of English-language works by Chinua Achebe and Wole Soyinka. The central question is not whether the translation of this or that work is “source” or “target” oriented, but to analyze how it manifests the “untranslatability” of the “literary decentering” (Meschonnic). We will analyze how these translations highlight the particularities and complexities of these texts, while providing their readers with a rich perspective on various linguistic and cultural aspects. Finally, we will demonstrate how the notion of “littérature-monde” gives visibility to these literatures in a context where they have historically been marginalized.Translation studies, Translation, Decentering, Untranslatable, Littérature-monde, Literary, Theories, Postcolonialism, Errors, cultures.De nombreuses études soulignent les « erreurs » présentes dans les traductions des œuvres de la littérature africaine, « erreurs » souvent attribuées à une connaissance insuffisante des contextes linguistiques, géo-historiques et culturels. Elles se manifestent notamment par des modifications sémantiques injustifiées, des traductions approximatives et des contresens. Toutefois, ces problèmes ne sont pas spécifiques à la littérature africaine. En réalité, toute œuvre littéraire, quelle que soit son origine, requiert une traduction attentive, capable d’en restituer toutes les nuances et dimensions. La plupart de ces études partent du principe que la littérature africaine est traduisible au sens traditionnel du terme. Nous adopterons, en revanche, la perspective selon laquelle toute œuvre est, par définition, « intraduisible », comme le soutient notamment Barbara Cassin. L’original n’est jamais entièrement traduisible en raison du passage inévitable d’une langue à une autre, qui fait partie intégrante du processus de traduction. Cela ne signifie pas pour autant que la traduction soit nécessairement inférieure à l’original. Au contraire, comme le démontre Borges dans Les traductions homériques et comme le soulignent des analyses récentes de Gérard Genette, l’œuvre doit être envisagée non seulement dans sa version originale, mais également à travers ses diverses traductions, qui représentent autant de réinterprétations possibles de celle-ci.Dès lors, la question ne se limite plus à la recherche de la traduction parfaite. Même si l’on suppose, notamment dans le cas de l’auto-traduction, qu’aucune « erreur » au sens traditionnel n’est commise, il n’en demeure pas moins que la meilleure des traductions sera toujours différente de l’original. C’est précisément ce qui donne toute sa valeur à la traduction : elle met en lumière des aspects de l’œuvre qui ne peuvent s’exprimer que grâce à elle, alors que l’original ne peut les contenir que de manière latente. La notion d’« erreur » au sens traditionnel perd de sa pertinence lorsque ces traductions s’inscrivent dans la Weltliteratur de Goethe ou dans la Littérature-monde, telle que définie par Alain Mabanckou, qui englobe des littératures marginalisées marquées par une hybridité linguistique. C’est à la lumière de ces considérations que nous réexaminerons la question des « erreurs » de traduction, en adoptant une approche problématisante plutôt qu’en cherchant à en établir une définition stricte.Notre analyse se fondera sur les traductions en langue anglaise de deux œuvres francophones de l’auteur ivoirien Ahmadou Kourouma, ainsi que sur les traductions françaises des œuvres anglophones de Chinua Achebe et de Wole Soyinka. Notre réflexion ne porte pas tant sur l’orientation « sourcière » ou « cibliste » d’une œuvre traduite que sur sa capacité à restituer l’« intraduisibilité » inhérente au décentrement de l’écriture (Meschonnic). Nous analyserons comment ces traductions mettent en lumière les particularités et les complexités de ces textes, tout en offrant à leurs lecteurs une perspective enrichie sur divers aspects langagiers et culturels. Nous veillerons enfin à démontrer comment la notion de « littérature-monde » contribue à valoriser ces littératures, leur conférant une visibilité essentielle dans un contexte où elles ont historiquement été marginalisées.Traductologie, Traduction, Décentrement, Intraduisibles, Littérature-monde, Littéraire, Théories, Post-colonialisme, Erreurs, Cultures
Theoretical Framework for Semitransparent Organic Photovoltaics: Bridging Material Design with Optical Engineering
No full textInternational audienceSemitransparent organic photovoltaics (ST‐OPVs) hold great promise for building‐integrated applications but encounter a critical trade‐off between power conversion efficiency (PCE) and visible transparency. This work addresses this challenge by introducing a novel theoretical framework that synergistically combines optical modeling and material design. A piecewise Gaussian absorption model has been developed to accurately simulate asymmetric spectral characteristics of organic materials. This model enables the virtual screening of potential organic donor‐acceptor configurations, identifying two optimal configurations: a narrow‐narrow (N + N) system and a wide‐narrow (W + N) system, which achieve a superior balance between PCE and transparency. High‐throughput optical screening of bilayer coupling layers further enhances light utilization efficiency to 9.8% (N + N) and 9.3% (W + N) through optimized light management. This study bridges material properties and optical engineering, providing a predictive tool for ST‐OPV optimization to guide the development of high‐performance, visually transparent solar technologies
Mamba Adaptive Anomaly Transformer with association discrepancy for time series
No full textInternational audienceAnomaly detection in time series poses a critical challenge in industrial monitoring, environmental sensing, and infrastructure reliability, where accurately distinguishing anomalies from complex temporal patterns remains an open problem. While existing methods, such as the Anomaly Transformer leveraging multi-layer association discrepancy between prior and series distributions and Dual Attention Contrastive Representation Learning architecture (DCdetector) employing dual-attention contrastive learning, have advanced the field, critical limitations persist. These include sensitivity to short-term context windows, computational inefficiency, and degraded performance under noisy and non-stationary real-world conditions. To address these challenges, we present MAAT (Mamba Adaptive Anomaly Transformer), an enhanced architecture that refines association discrepancy modeling and reconstruction quality for more robust anomaly detection. Our work introduces two key contributions to the existing Anomaly transformer architecture: Sparse Attention, which computes association discrepancy more efficiently by selectively focusing on the most relevant time steps. This reduces computational redundancy while effectively capturing long-range dependencies critical for discerning subtle anomalies. A Mamba-Selective State Space Model (Mamba-SSM) is also integrated into the reconstruction module. A skip connection bridges the original reconstruction and the Mamba-SSM output, while a Gated Attention mechanism adaptively fuses features from both pathways. This design balances fidelity and contextual enhancement dynamically, improving anomaly localization and overall detection performance. Extensive experiments on benchmark datasets demonstrate that MAAT significantly outperforms prior methods, achieving superior anomaly distinguishability and generalization across diverse time series applications. By addressing the limitations of existing approaches, MAAT sets a new standard for unsupervised time series anomaly detection in real-world scenarios. Code available at https://github.com/ilyesbenaissa/MAAT.</div