Search

Your search keyword '"A. Baert"' showing total 19,572 results
Start Over Author "A. Baert"

Refine your results

more »

more »

more »

more »

more »

more »

more »
19,572 results on '"A. Baert"'

1. Reward Machine Inference for Robotic Manipulation

Author

Baert, Mattijs, Leroux, Sam, and Simoens, Pieter

Subjects
Computer Science - Robotics, Computer Science - Machine Learning
Abstract

Learning from Demonstrations (LfD) and Reinforcement Learning (RL) have enabled robot agents to accomplish complex tasks. Reward Machines (RMs) enhance RL's capability to train policies over extended time horizons by structuring high-level task information. In this work, we introduce a novel LfD approach for learning RMs directly from visual demonstrations of robotic manipulation tasks. Unlike previous methods, our approach requires no predefined propositions or prior knowledge of the underlying sparse reward signals. Instead, it jointly learns the RM structure and identifies key high-level events that drive transitions between RM states. We validate our method on vision-based manipulation tasks, showing that the inferred RM accurately captures task structure and enables an RL agent to effectively learn an optimal policy.

Published
2024

2. Bluetooth Low Energy Dataset Using In-Phase and Quadrature Samples for Indoor Localization

Author

Leitch, Samuel G., Ahmed, Qasim Zeeshan, Van Herbruggen, Ben, Baert, Mathias, Fontaine, Jaron, De Poorter, Eli, Shahid, Adnan, and Lazaridis, Pavlos I.

Subjects
Computer Science - Machine Learning, Electrical Engineering and Systems Science - Signal Processing
Abstract

One significant challenge in research is to collect a large amount of data and learn the underlying relationship between the input and the output variables. This paper outlines the process of collecting and validating a dataset designed to determine the angle of arrival (AoA) using Bluetooth low energy (BLE) technology. The data, collected in a laboratory setting, is intended to approximate real-world industrial scenarios. This paper discusses the data collection process, the structure of the dataset, and the methodology adopted for automating sample labeling for supervised learning. The collected samples and the process of generating ground truth (GT) labels were validated using the Texas Instruments (TI) phase difference of arrival (PDoA) implementation on the data, yielding a mean absolute error (MAE) at one of the heights without obstacles of $25.71^\circ$. The distance estimation on BLE was implemented using a Gaussian Process Regression algorithm, yielding an MAE of $0.174$m.

Published
2024

3. SPARK: Self-supervised Personalized Real-time Monocular Face Capture

Author

Baert, Kelian, Bharadwaj, Shrisha, Castan, Fabien, Maujean, Benoit, Christie, Marc, Abrevaya, Victoria, and Boukhayma, Adnane

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning
Abstract

Feedforward monocular face capture methods seek to reconstruct posed faces from a single image of a person. Current state of the art approaches have the ability to regress parametric 3D face models in real-time across a wide range of identities, lighting conditions and poses by leveraging large image datasets of human faces. These methods however suffer from clear limitations in that the underlying parametric face model only provides a coarse estimation of the face shape, thereby limiting their practical applicability in tasks that require precise 3D reconstruction (aging, face swapping, digital make-up, ...). In this paper, we propose a method for high-precision 3D face capture taking advantage of a collection of unconstrained videos of a subject as prior information. Our proposal builds on a two stage approach. We start with the reconstruction of a detailed 3D face avatar of the person, capturing both precise geometry and appearance from a collection of videos. We then use the encoder from a pre-trained monocular face reconstruction method, substituting its decoder with our personalized model, and proceed with transfer learning on the video collection. Using our pre-estimated image formation model, we obtain a more precise self-supervision objective, enabling improved expression and pose alignment. This results in a trained encoder capable of efficiently regressing pose and expression parameters in real-time from previously unseen images, which combined with our personalized geometry model yields more accurate and high fidelity mesh inference. Through extensive qualitative and quantitative evaluation, we showcase the superiority of our final model as compared to state-of-the-art baselines, and demonstrate its generalization ability to unseen pose, expression and lighting., Comment: SIGGRAPH Asia 2024 Conference Paper. Project page: https://kelianb.github.io/SPARK/

Published
2024
Full Text
View/download PDF

4. Learning Task Specifications from Demonstrations as Probabilistic Automata

Author

Baert, Mattijs, Leroux, Sam, and Simoens, Pieter

Subjects
Computer Science - Robotics
Abstract

Specifying tasks for robotic systems traditionally requires coding expertise, deep domain knowledge, and significant time investment. While learning from demonstration offers a promising alternative, existing methods often struggle with tasks of longer horizons. To address this limitation, we introduce a computationally efficient approach for learning probabilistic deterministic finite automata (PDFA) that capture task structures and expert preferences directly from demonstrations. Our approach infers sub-goals and their temporal dependencies, producing an interpretable task specification that domain experts can easily understand and adjust. We validate our method through experiments involving object manipulation tasks, showcasing how our method enables a robot arm to effectively replicate diverse expert strategies while adapting to changing conditions.

Published
2024

5. Exploring the Perception and Retention of Movement Analysis Skills through Online Mastery-Based Modules

Author

Helena Baert, Larissa True, Matthew Madden, Erica Pratt, and Emily Gilbert

Abstract

A mixed-methods study was used to determine if an online, mastery-based learning resource was successful in improving content knowledge and retention of critical elements and developmental stages of 16 fundamental movement skills (FMS); a secondary purpose was to determine participant perceptions of their experience with the online resource. Participants (N = 323; motor development (MD) = 94; movement education (ME) = 124; elementary physical education (EPE) = 105) were Teacher Candidates (TCs) in a Physical Education Teacher Education (PETE) Program. Pre- and post-survey data, and pre-, post-, and retention test scores were used to evaluate the program's effectiveness. Paired samples t-tests indicated higher post-test scores for MD ([delta]37.23, p < 0.001, d = 2.32) and ME ([delta]32.12, p < 0.001, d = 1.72). Results from a one-way ANOVA indicated significant improvement and retention of test scores over time, F(3, 331) = 27.761, p < 0.001, [eta-squared] = 0.963. Perceptions of PETE TCs reported positive reactions to the use of the online modules to improve their ability to analyze FMS. Based on these findings, the authors propose that implementing an online, mastery-based resource to analyze FMS may be a positive learning experience for TCs and could lead to long-term content knowledge acquisition of critical elements of FMS.

Published
2024

6. Is the EJRA proportionate and therefore justified? A critical review of the EJRA policy at Cambridge

Author

Linton, Oliver, Rau, Raghavendra, Baert, Patrick, Bossaerts, Peter, Crowcroft, Jon, Evans, G. R., Ewart, Paul, Gay, Nick, Kattuman, Paul, Scholtes, Stefan, Sabourian, Hamid, and Smith, Richard J.

Subjects
Economics - General Economics
Abstract

This paper critically evaluates the HESA (Higher Education Statistics Agency) Data Report for the Employer Justified Retirement Age (EJRA) Review Group at the University of Cambridge (\cite{CambridgeHESA2024}), identifying significant methodological flaws and misinterpretations. Our analysis reveals issues such as unclear application of data filters, inconsistent variable treatment, and erroneous statistical conclusions. The Report suggests that the EJRA increased job creation rates at Cambridge, but we show Cambridge consistently had lower job creation rates for Established Academic Careers compared to other Russell Group universities, both before and after EJRA implementation in 2011, with no evidence for a significant change in this deficit post implementation. This suggests that EJRA is not a significant factor driving job creation rates. Since other universities without an EJRA exhibit higher job creation rates, this suggests job creation can be sustained without such a policy. We conclude that the EJRA did not achieve its intended goal of increasing opportunities for young academics and may have exacerbated existing disparities compared to other leading universities. We recommend EJRA be abolished at Cambridge since it does not meet its aims and could be viewed as unlawful age discrimination.\newline \it{This version is a revision reflecting some of the comments made by members of the university EJRA review group in public discussion, see \cite{Holmes24}.}, Comment: This is a revision of our earlier paper

Published
2024

7. Report of the sixth meeting of the European Consortium ‘Care for CMMRD’ (C4CMMRD), Paris, France, November 16th 2022

Author

Guerrini-Rousseau, Léa, Gallon, Richard, Pineda, Marta, Brugières, Laurence, Baert-Desurmont, Stéphanie, Corsini, Carole, Dangouloff-Ros, Volodia, Gorris, Mark A. J., Haberler, Christine, Hoarau, Pauline, Jongmans, Marjolijn C., Kloor, Matthias, Loeffen, Jan, Rigaud, Charlotte, Robbe, Julie, Vibert, Roseline, Weijers, Dilys, Wimmer, Katharina, and Colas, Chrystelle

Published
2024
Full Text
View/download PDF

10. Learning Safety Constraints From Demonstration Using One-Class Decision Trees

Author

Baert, Mattijs, Leroux, Sam, and Simoens, Pieter

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence
Abstract

The alignment of autonomous agents with human values is a pivotal challenge when deploying these agents within physical environments, where safety is an important concern. However, defining the agent's objective as a reward and/or cost function is inherently complex and prone to human errors. In response to this challenge, we present a novel approach that leverages one-class decision trees to facilitate learning from expert demonstrations. These decision trees provide a foundation for representing a set of constraints pertinent to the given environment as a logical formula in disjunctive normal form. The learned constraints are subsequently employed within an oracle constrained reinforcement learning framework, enabling the acquisition of a safe policy. In contrast to other methods, our approach offers an interpretable representation of the constraints, a vital feature in safety-critical environments. To validate the effectiveness of our proposed method, we conduct experiments in synthetic benchmark domains and a realistic driving environment., Comment: accepted for AAAI 2024 Workshop on Neuro-Symbolic Learning and Reasoning in the Era of Large Language Models (NucLeaR)

Published
2023

12. A self-amplifying RNA RSV prefusion-F vaccine elicits potent immunity in pre-exposed and naïve non-human primates

Author

Vijayan, Aneesh, Vogels, Ronald, Groppo, Rachel, Jin, Yi, Khan, Selina, Van Kampen, Mirjam, Jorritsma, Sytze, Boedhoe, Satish, Baert, Miranda, van Diepen, Harry, Kuipers, Harmjan, Serroyen, Jan, del Valle, Jorge Reyes-, Broman, Ann, Nguyen, Lannie, Ray, Sayoni, Jarai, Bader, Arora, Jayant, Lifton, Michelle, Mildenberg, Benjamin, Morton, Georgeanna, Santra, Sampa, Grossman, Tamar R., Schuitemaker, Hanneke, Custers, Jerome, and Zahn, Roland

Published
2024
Full Text
View/download PDF

15. Cytomegalovirus infection during pregnancy: cross-sectional survey of knowledge and prevention practices of healthcare professionals in French-speaking Switzerland

Author

Sartori, Pauline, Baud, David, de Tejada, Begoña Martinez, Farin, Alexandre, Rossier, Marie-Claude, Rieder, Wawrzyniec, Rouiller, Sylvie, Robyr, Romaine, Grant, Gaston, Eggel, Béatrice, Bruno, Adrien, Pretalli, Maeva, Gainon, Bertrand, Capoccia-Brugger, Romina, Ammon-Chansel, Laetitia, De Courten, Clarisse, Beurret-Lepori, Nathalie, Baert, Jonathan, Vial, Yvan, and Pomar, Léo

Published
2024
Full Text
View/download PDF

16. Time Tetris: a longitudinal study on compressed schedules and workplace well-being at IKEA

Author

Kristen du Bois, Stijn Baert, Louis Lippens, and Eva Derous

Subjects
Compressed schedule, Psychological detachment, Exhaustion, Burnout, Well-being, Public aspects of medicine, RA1-1270
Abstract

Abstract Background Compressed schedules, where workers perform longer daily hours to enjoy additional days off, are increasingly promoted as a workplace well-being intervention. Nevertheless, their implications for work-related well-being outcomes, such as recovery from work and burnout risk, are understudied. This gap leaves employers with little evidence on whether and how the arrangement contributes to workplace well-being. Methods IKEA Belgium offered its employees the option to enter compressed schedules in the aftermath of a national labour reform aimed at improving well-being and reducing burnout. We collected data on psychological detachment from work, work-related exhaustion, and burnout risk in four waves before and after implementation. We used mixed-effects growth models to estimate the within-subjects changes in these three domains, and two-way fixed effects models to compare changes with those from a non-treated comparison group. Results Workers experienced increased psychological detachment from work in compressed schedules, yet we saw no decrease in work-related exhaustion or burnout risk. While between-subjects analyses confirm that the increase in psychological detachment is related to treatment, they also hint that this association may fade out during summer when all workers take more extended breaks from work. Conclusions While workers in compressed schedules may mentally switch off from work more effectively, this does not translate into decreased burnout risk scores. Consistent with theoretical expectations, policymakers and employers should be cautious in assuming that the arrangements significantly reduce burnout.

Published
2025
Full Text
View/download PDF

18. A versatile single-photon-based quantum computing platform

Author

Maring, Nicolas, Fyrillas, Andreas, Pont, Mathias, Ivanov, Edouard, Stepanov, Petr, Margaria, Nico, Hease, William, Pishchagin, Anton, Lemaître, Aristide, Sagnes, Isabelle, Au, Thi Huong, Boissier, Sébastien, Bertasi, Eric, Baert, Aurélien, Valdivia, Mario, Billard, Marie, Acar, Ozan, Brieussel, Alexandre, Mezher, Rawad, Wein, Stephen C., Salavrakos, Alexia, Sinnott, Patrick, Fioretto, Dario A., Emeriau, Pierre-Emmanuel, Belabas, Nadia, Mansfield, Shane, Senellart, Pascale, Senellart, Jean, and Somaschi, Niccolo

Published
2024
Full Text
View/download PDF

20. A general-purpose single-photon-based quantum computing platform

Author

Maring, Nicolas, Fyrillas, Andreas, Pont, Mathias, Ivanov, Edouard, Stepanov, Petr, Margaria, Nico, Hease, William, Pishchagin, Anton, Au, Thi Huong, Boissier, Sébastien, Bertasi, Eric, Baert, Aurélien, Valdivia, Mario, Billard, Marie, Acar, Ozan, Brieussel, Alexandre, Mezher, Rawad, Wein, Stephen C., Salavrakos, Alexia, Sinnott, Patrick, Fioretto, Dario A., Emeriau, Pierre-Emmanuel, Belabas, Nadia, Mansfield, Shane, Senellart, Pascale, Senellart, Jean, and Somaschi, Niccolo

Subjects
Quantum Physics
Abstract

Quantum computing aims at exploiting quantum phenomena to efficiently perform computations that are unfeasible even for the most powerful classical supercomputers. Among the promising technological approaches, photonic quantum computing offers the advantages of low decoherence, information processing with modest cryogenic requirements, and native integration with classical and quantum networks. To date, quantum computing demonstrations with light have implemented specific tasks with specialized hardware, notably Gaussian Boson Sampling which permitted quantum computational advantage to be reached. Here we report a first user-ready general-purpose quantum computing prototype based on single photons. The device comprises a high-efficiency quantum-dot single-photon source feeding a universal linear optical network on a reconfigurable chip for which hardware errors are compensated by a machine-learned transpilation process. Our full software stack allows remote control of the device to perform computations via logic gates or direct photonic operations. For gate-based computation we benchmark one-, two- and three-qubit gates with state-of-the art fidelities of $99.6\pm0.1 \%$, $93.8\pm0.6 \%$ and $86\pm1.2 \%$ respectively. We also implement a variational quantum eigensolver, which we use to calculate the energy levels of the hydrogen molecule with high accuracy. For photon native computation, we implement a classifier algorithm using a $3$-photon-based quantum neural network and report a first $6$-photon Boson Sampling demonstration on a universal reconfigurable integrated circuit. Finally, we report on a first heralded 3-photon entanglement generation, a key milestone toward measurement-based quantum computing.

Published
2023

21. The key to bringing DNA collections to the next level

Author

Emily Veltjen, Pieter Asselman, Wim Baert, Steve Baeyen, Lise Beirinckx, Liselot Breyne, Dimitri Brosens, Tim Claerhout, Sari Cogneau, Karen Cox, Laura Cuypers, Lynn Delgat, Philippe Desmeth, Jordi de Raad, Lore Esselens, Maria-Rose Eves Down, Philippe Helsen, Frederik Leliaert, Kenny Meganck, Zjef Pereboom, Nathalie Smitz, Gontran Sonet, Maarten Trekels, An Vanden Broeck, Charlotte Van Driessche, and Aaike De Wever

Subjects
DNA specimens, Natural Science collections, DiSSCo, Science
Abstract

DNA collections are a valuable type of Natural Science collection, enabling the validation of past research, serving as a source for new genomic studies and supporting ex situ conservation. The DiSSCo Flanders DNA collection working group, aiming to advance and "unlock" their DNA collections, identified the need for: 1) actively sharing best practices regarding the management of DNA collections; and 2) providing guidance on how to bring theory into practice. By combining best practice examples from within the working group with available literature and brainstorming ideas, the working group co-created two outputs, referred to as: the "Challenges" and the "Key". The Challenges are a list of obstacles to DNA collection management, which shape the structure of the linked Key and can also be used to spark discussion amongst stakeholders. The Key is a tool that guides users through the maturation process of their DNA collection in a standardised way. It stimulates holistic growth, breaks down the needed work into manageable steps and helps to decide priorities during the process. Furthermore, the Key facilitates communication with both internal stakeholders and external DNA collection managers. The Key distinguishes itself from other self-assessment tools in several ways: it includes (re)investigation of the collection’s purpose and context; it is specialised for DNA collections; it delivers concrete goals linked to relevant information and shared experience; and it is inclusive, targeting all Natural Science DNA collections, regardless of their context or size.

Published
2024
Full Text
View/download PDF

22. Estimation of flea beetle damage in the field using a multistage deep learning-based solution

Author

Arantza Bereciartua-Pérez, María Monzón, Daniel Múgica, Greta De Both, Jeroen Baert, Brittany Hedges, Nicole Fox, Jone Echazarra, and Ramón Navarra-Mestre

Subjects
Convolutional neural networks, Deep learning, Plant phenotyping, Damage estimation, Plant crop detection and identification, Agriculture
Abstract

Estimation of damage in plants is a key issue for crop protection. Currently, experts in the field manually assess the plots. This is a time-consuming task that can be automated thanks to the latest technology in computer vision (CV). The use of image-based systems and recently deep learning-based systems have provided good results in several agricultural applications. These image-based applications outperform expert evaluation in controlled environments, and now they are being progressively included in non-controlled field applications.A novel solution based on deep learning techniques in combination with image processing methods is proposed to tackle the estimate of plant damage in the field. The proposed solution is a two-stage algorithm. In a first stage, the single plants in the plots are detected by an object detection YOLO based model. Then a regression model is applied to estimate the damage of each individual plant. The solution has been developed and validated in oilseed rape plants to estimate the damage caused by flea beetle.The crop detection model achieves a mean precision average of 91% with a mAP@0.50 of 0.99 and a mAP@0.95 of 0.91 for oilseed rape specifically. The regression model to estimate up to 60% of damage degree in single plants achieves a MAE of 7.11, and R2 of 0.46 in comparison with manual evaluations done plant by plant by experts. Models are deployed in a docker, and with a REST API communication protocol they can be inferred directly for images acquired in the field from a mobile device.

Published
2024
Full Text
View/download PDF

23. Maximum Causal Entropy Inverse Constrained Reinforcement Learning

Author

Baert, Mattijs, Mazzaglia, Pietro, Leroux, Sam, and Simoens, Pieter

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence
Abstract

When deploying artificial agents in real-world environments where they interact with humans, it is crucial that their behavior is aligned with the values, social norms or other requirements of that environment. However, many environments have implicit constraints that are difficult to specify and transfer to a learning agent. To address this challenge, we propose a novel method that utilizes the principle of maximum causal entropy to learn constraints and an optimal policy that adheres to these constraints, using demonstrations of agents that abide by the constraints. We prove convergence in a tabular setting and provide an approximation which scales to complex environments. We evaluate the effectiveness of the learned policy by assessing the reward received and the number of constraint violations, and we evaluate the learned cost function based on its transferability to other agents. Our method has been shown to outperform state-of-the-art approaches across a variety of tasks and environments, and it is able to handle problems with stochastic dynamics and a continuous state-action space.

Published
2023

24. Bursts from Space: MeerKAT - The first citizen science project dedicated to commensal radio transients

Author

Andersson, Alex, Lintott, Chris, Fender, Rob, Bright, Joe, Carotenuto, Francesco, Driessen, Laura, Espinasse, Mathilde, Gaseahalwe, Kelebogile, Heywood, Ian, van der Horst, Alexander J., Motta, Sara, Rhodes, Lauren, Tremou, Evangelia, Williams, David R. A., Woudt, Patrick, Zhang, Xian, Bloemen, Steven, Groot, Paul, Vreeswijk, Paul, Giarratana, Stefano, Saikia, Payaswini, Andersson, Jonas, Arroyo, Lizzeth Ruiz, Baert, Loïc, Baumann, Matthew, Domainko, Wilfried, Eschweiler, Thorsten, Forsythe, Tim, Gaudenzi, Sauro, Grenier, Rachel Ann, Iannone, Davide, Lahoz, Karla, Melville, Kyle J., Nascimento, Marianne De Sousa, Navarro, Leticia, Parthasarathi, Sai, Piilonen, Rahman, Najma, Smith, Jeffrey, Stewart, B., Temoke, Newton, Tworek, Chloe, and Whittle, Isabelle

Subjects
Astrophysics - High Energy Astrophysical Phenomena, Astrophysics - Astrophysics of Galaxies, Astrophysics - Instrumentation and Methods for Astrophysics
Abstract

The newest generation of radio telescopes are able to survey large areas with high sensitivity and cadence, producing data volumes that require new methods to better understand the transient sky. Here we describe the results from the first citizen science project dedicated to commensal radio transients, using data from the MeerKAT telescope with weekly cadence. Bursts from Space: MeerKAT was launched late in 2021 and received ~89000 classifications from over 1000 volunteers in 3 months. Our volunteers discovered 142 new variable sources which, along with the known transients in our fields, allowed us to estimate that at least 2.1 per cent of radio sources are varying at 1.28 GHz at the sampled cadence and sensitivity, in line with previous work. We provide the full catalogue of these sources, the largest of candidate radio variables to date. Transient sources found with archival counterparts include a pulsar (B1845-01) and an OH maser star (OH 30.1-0.7), in addition to the recovery of known stellar flares and X-ray binary jets in our observations. Data from the MeerLICHT optical telescope, along with estimates of long time-scale variability induced by scintillation, imply that the majority of the new variables are active galactic nuclei. This tells us that citizen scientists can discover phenomena varying on time-scales from weeks to several years. The success both in terms of volunteer engagement and scientific merit warrants the continued development of the project, whilst we use the classifications from volunteers to develop machine learning techniques for finding transients., Comment: Accepted to MNRAS, 14 pages + an appendix containing our main data table

Published
2023
Full Text
View/download PDF

25. Exploring Crossmodal Interaction of Tactile and Visual Cues on Temperature Perception in Virtual Reality: a Preliminary Study

Author

Helfenstein-Didier, Clémentine, Dhouib, Amira, Favre, Florent, Pascal, Jonathan, and Baert, Patrick

Subjects
Computer Science - Human-Computer Interaction
Abstract

VEs are typically limited to visual and auditory cues; however, recent results show that multiple sensory modalities increase the immersion. In this study, an experimental protocol is proposed to recreate multiple tactile, in particular thermal, sensations in VR. The aim is twofold: (1) studying the performance of different devices for creating warm and cold sensations with regards to their efficiency and acoustic disturbance; and (2) investigating the interdependency between visual and tactile stimuli in the perception of temperature. 14 participants performed two experimental studies. Our results show no acoustic disturbance of the materials used. Spot projector is more efficient than fan heater to create a warm sensation; fan + water spray is more efficient than fan alone to create cold sensation. Moreover, no significant contribution of visual cue on the thermal perception was found except for the extremely cold simulation (snow visualization and thermal stimulation performed with fan + water spray).

Published
2023

26. Partial rejuvenation of the spermatogonial stem cell niche after gender-affirming hormone therapy in trans women

Author

Emily Delgouffe, Samuel Madureira Silva, Frédéric Chalmel, Wilfried Cools, Camille Raets, Kelly Tilleman, Guy T'Sjoen, Yoni Baert, and Ellen Goossens

Subjects
transgender women, spermatogonial stem cells, testicular tissue, spermatogenesis, hormonal treatment, germ cells, Medicine, Science, Biology (General), QH301-705.5
Abstract

Although the impact of gender-affirming hormone therapy (GAHT) on spermatogenesis in trans women has already been studied, data on its precise effects on the testicular environment is poor. Therefore, this study aimed to characterize, through histological and transcriptomic analysis, the spermatogonial stem cell niche of 106 trans women who underwent standardized GAHT, comprising estrogens and cyproterone acetate. A partial dedifferentiation of Sertoli cells was observed, marked by the co-expression of androgen receptor and anti-Müllerian hormone which mirrors the situation in peripubertal boys. The Leydig cells also exhibited a distribution analogous to peripubertal tissue, accompanied by a reduced insulin-like factor 3 expression. Although most peritubular myoid cells expressed alpha-smooth muscle actin 2, the expression pattern was disturbed. Besides this, fibrosis was particularly evident in the tubular wall and the lumen was collapsing in most participants. A spermatogenic arrest was also observed in all participants. The transcriptomic profile of transgender tissue confirmed a loss of mature characteristics - a partial rejuvenation - of the spermatogonial stem cell niche and, in addition, detected inflammation processes occurring in the samples. The present study shows that GAHT changes the spermatogonial stem cell niche by partially rejuvenating the somatic cells and inducing fibrotic processes. These findings are important to further understand how estrogens and testosterone suppression affect the testis environment, and in the case of orchidectomized testes as medical waste material, their potential use in research.

Published
2025
Full Text
View/download PDF

27. Changes in serum concentration of perioperative inflammatory cytokines following the timing of surgery among mild–moderate traumatic brain injury patients and factors associated

Author

Hervé Monka Lekuya, Stephen Cose, Marjorie Nakibuule, Gift Ahimbisibwe, Anthony Fuller, Larrey Kasereka Kamabu, Emmanuel Biryabarema, Geoffrey Olweny, David Patrick Kateete, Anthony Kirabira, Fredrick Makumbi, Jelle Vandersteene, Edward Baert, Moses Galukande, and Jean-Pierre Okito Kalala

Subjects
neuro-inflammation, perioperative serum levels inflammatory cytokines, systemic inflammatory response, timing of surgery, traumatic brain injury, Neurology. Diseases of the nervous system, RC346-429
Abstract

BackgroundThe safe timing window for surgery during the acute phase of inflammation due to traumatic brain injury (TBI) has not been studied extensively. We aimed to elucidate the relationship between the timing of surgery and changes in perioperative serum levels of inflammatory cytokines and factors associated to optimize TBI management in low-middle-income countries.MethodsA prospective cohort study was conducted among TBI Patients with depressed skull fractures with a GCS > 8 operated at different timing from injury and followed up peri-operatively. We collected the clinical-radiological data, as well as pre-and postoperative venous samples from participants; we then did Luminex Assay to quantify the serum levels of pro/anti-inflammatory cytokines using the kits of 96-well human cytokine “27-Plex-Assay (#M500KCAF0Y®).” We performed the analysis with STATA version 17 and R_studio applying both descriptive and inferential methods.ResultsWe enrolled 82 TBI patients with a median (IQR) age of 25.5 (20–34) years, and the majority were male (85.4%). There were 48.8% victims of assaults, and 73.2% had a post-resuscitation admission GCS of 14–15. There were 38 (46.3%) who were operated within 48 h of injury versus 44 (53.7%) after 48 h. Serum levels of TNF-α were significantly higher after surgeries done >48 h compared to those done ≤48 h (p = 0.0327); whereas, the difference in post-operative mean serum levels of IL-10 was significantly increased in patients who developed later SSI compared to those who did not (11.56 versus −0.58 pg./mL, p = 0.0489). In multivariate analysis, the history of post-traumatic seizure (PTS) was associated with a postoperative increase in TNF-α (p = 0.01), the hemoglobin of 10–12 with a postoperative decrease of IL-4 (p = 0.05); the presence of focal neurological deficit was associated with a significant postoperative increased of TNF-α, IL-6, and IL-4 (p = 0.05). The presence of extra-axial hemorrhage was associated with a postoperative increase of IL-10 (p = 0.05).ConclusionDelayed surgical intervention beyond 48 h post-injury in mild–moderate TBI patients results in a significantly increased postoperative inflammatory response, as evidenced by elevated serum levels of TNF-α and IL-6. Neurological deficits, PTS, reduced hemoglobin rate, and extra-axial intracranial hemorrhage are factors associated with this heightened response.

Published
2024
Full Text
View/download PDF

28. Comparison of the ADNEX and ROMA risk prediction models for the diagnosis of ovarian cancer: a multicentre external validation in patients who underwent surgery

Author

Landolfo, Chiara, Ceusters, Jolien, Valentin, Lil, Froyman, Wouter, Van Gorp, Toon, Heremans, Ruben, Baert, Thaïs, Wouters, Roxanne, Vankerckhoven, Ann, Van Rompuy, Anne-Sophie, Billen, Jaak, Moro, Francesca, Mascilini, Floriana, Neumann, Adam, Van Holsbeke, Caroline, Chiappa, Valentina, Bourne, Tom, Fischerova, Daniela, Testa, Antonia, Coosemans, An, Timmerman, Dirk, and Van Calster, Ben

Published
2024
Full Text
View/download PDF

36. Guselkumab in patients with moderately to severely active ulcerative colitis (QUASAR): phase 3 double-blind, randomised, placebo-controlled induction and maintenance studies

Author

Abrahamovych, Orest, Abu-farsakh, Niazy, Afanasieva, Halyna, Akpinar, Hale, Al Kharrat, Houssam, Altintas, Engin, Altwegg, Romain, Andreev, Pavel, Aomatsu, Kazuki, Araki, Hiroshi, Argollo, Marjorie, Ariel, Federico, Armuzzi, Alessandro, Ashida, Toshifumi, Augustyn, Monika, Aumais, Guy, Azum, Martín, Baert, Filip, Balaz, Jozef, Balderramo, Domingo, Begun, Jakob, Berova, Temenuzhka, Billiauws, Lore, Blanco, Antonio, Bortlik, Martin, Bossa, Fabrizio, Bunkova, Elena, Cabello, Mercedes, Cao, Qian, Caprioli, Flavio, Cerqueira, Rute, Chachu, Karen, Chamouard, Patrick, Chen, Chunxiao, Chen, Yan, Chen, Baili, Chen, Dongfeng, Chen, Hong, Chen, Youxiang, Chen, Chou-chen, Cheon, Jaehee, Chen, Minhu, Chiu, Cheng-tang, Choi, Changhwan, Chow, Elizabeth, Cicala, Michele, Cintra, Aderson, Danilkiewicz, Wit, Datsenko, Olena, De Hertogh, Gert, De María, Julio, Del Valle Torrealba Medina, Leyanira, Deliang, Liu, Desreumaux, Pierre, Dewint, Pieter, Ding, Shigang, Doherty, Glen, Draganova, Raina, Dutré, Joris, Duvall, George, Egan, Laurence, Fahed, Julien, Fechner, Lars, Fedurco, Miroslav, Ferhat Celik, Aykut, Fernandez, Juan, Filip, Rafal, Fishman, Sigal, Flores, Cristina, Fogel, Ronald, Fowler, Sharyle, Francesconi, Carlos, Fujii, Toshimitsu, Fujiya, Mikihiro, Fukata, Masayuki, Fukuhara, Seiichiro, Furumoto, Yohei, G.Kiss, Gyula, Gachowski, Waldemar, Gao, Xiang, Gasbarrini, Antonio, Gawdis-wojnarska, Beata, Gaya, Daniel, Geccherle, Andrea, Gilletta De Saint Joseph, Cyrielle, Gimenez, Edgardo, Gionchetti, Paolo, Goldin, Eran, Golovchenko, Oleksandr, Gonciarz, Maciej, Gonen, Can, Gonzales, Chad, Gonzalez, Raquel, Gonzalez Segura, Gaston Julian, Gridnyev, Oleksiy, Gu, Fang, Gupta, Nitin, Gyokeres, Tibor, Haines, Melissa, Hebuterne, Xavier, Hedin, Charlotte, Hellstrom, Per, Hilmi, Ida Normiha, Hirai, Fumihito, Hisabe, Takashi, Hlavaty, Tibor, Hoentjen, Frank, Holtmann, Gerald, Horiki, Noriyuki, Horny, Ivo, Horvat, Gyula, Horvath, Frantisek, Hoshi, Namiko, Hosomi, Shuhei, Hospodarskyy, Ihor, Hou, Xiaohua, Hrdlicka, Ludek, Hu, Naizhong, Huang, Vivian, Huang, Caibin, Inaba, Tomoki, Ishiguro, Yoh, Ishihara, Shunji, Ishikawa, Takeshi, Iskandar, Heba, Israeli, Eran, Itaba, Soichi, Ivanishyn, Olha, Izbéki, Ferenc, Jain, Animesh, Jairath, Vipul, Jang, Byung Ik, Jiang, Chunmeng, Jones, Michael, Kagaya, Takashi, Kalimullina, Dilara, Karakina, Marina, Karatas, Ali, Kashimura, Hiroshi, Keret, Dan, Khovaeva, Yaroslava, Kierkus, Jaroslaw, Kim, Taeoh, Kim, Youngho, Kim, Hyo Jong, Klaus, Jochen, Kleczkowski, Dariusz, Klopocka, Maria, Kobayashi, Taku, Kobielusz-gembala, Iwona, Kojima, Yuichiro, Koo, Ja Seol, Kopon, Adam, Kopp, James, Koyama, Fumikazu, Koynarski, Vasil, Kravchenko, Tetiana, Kudo, Masatoshi, Kusunoki, Ryusaku, Kwon, Kwangan, Laclav, Martin, Lago, Paula, Laharie, David, Lasa, Juan, Lateef, Syed, Latinovic Radakovic, Tatjana, Lee, Kang Moon, Leszczyszyn, Jaroslaw, Li, Yan, Liu, Mei, Liu, Eashen, Lobaton Ortega, Triana, Lopatin, Denis, Lopes, Susana, Lovric Jovanovic, Mileva, Lubini, Marcio, Lukas, Milan, Maaser, Christian, Machytka, Evzen, Maemoto, Atsuo, Magro, Fernando, Marcet, Jorge, Marini, Eduardo, Marusawa, Hiroyuki, Matous, Jan, Matsumoto, Takayuki, Mendu, Shoba, Miao, Yinglei, Mihaly, Emese, Miheller, Pal, Ministro, Paula, Miyabayashi, Hideharu, Mohl, Wolfgang, Molnar, Martin, Moore, Gregory, Moran Faienzo, Gabriela, Morral, Eugeni Domenech, Motoya, Satoshi, Murali, Narayanachar, Mustaffa, Nazri Mohamad, Naem, Mohamed, Nakai, Katsuhiko, Nakajima, Koichi, Nakamoto, Yasunari, Nakamura, Masanao, Nancey, Stephane, Navaneethan, Udayakumar, Nawawi, Khairul Najmi Muhammad, Ninomiya, Tomoyuki, Novak, Janos, Oki, Motoji, Oliveira Santana, Genoile, Onizawa, Michio, Ono, Yohei, Osada, Taro, Osipenko, Marina, Owczarek, Danuta, Parente, Jose Miguel, Patel, Kamal, Patel, Bhaktasharan, Pedersen, Peder, Petroniene, Rima, Petrov, Asen, Petrova, Elina, Piquet, Marie-astrid, Pokrotnieks, Juris, Poroshina, Elena, Portela, Francisco, Pruthi, Jatinder S., Prystupa, Lyudmila, Pukitis, Aldis, Qiu, Yun, Rafalsky, Vladimir, Rainis, Tova, Ramos Junior, Odery, Rashid, Mohammed, Rayyan, Yasser, Reshetko, Olga, Ribeiro, Tarsila, Rivero, Montserrat, Roblin, Xavier, Romatowski, Jacek, Rowbotham, David, Ruffinengo, Orlando Enrique, Rydzewska-wyszkowska, Grazyna, Sablin, Oleg, Sai, Souken, Saibeni, Simone, Said, Rosaida Hj. Md., Sakuraba, Hirotake, Samaan, Mark, Sandborn, William, Saruta, Masayuki, Sauk, Jenny, Savarino, Edoardo Vincenzo, Schultz, Michael, Schulze, Joerg, Sedghi, Shahriar, Seidler, Ursula, Seksik, Philippe, Senturk, Omer, Shelton, Edward, Shimizu, Masahito, Shin, Sungjae, Shukla, Richa, Sigal, Michael, Silva Junior, Roberto, Simoes Neto, Joaquim, Siroky Jr., Milan, Smajstrla, Vit, Smid, Vaclav, Smiley, Sarah, Soledad Brunatto, Luciana, Stanislavchuk, Mykola, Stokesberry, David, Stone, Christian, Suiter, Daniel, Suzuki, Takayoshi, Svorcan, Petar, Takagi, Sho, Takagi, Tomohisa, Takamura, Masaaki, Takayama, Tetsuji, Takeuchi, Ken, Tanaka, Akihito, Tanaka, Hiroki, Tanash, Omar, Tang, Wen, Tankova, Ludmila, Teixeira Campos, Luciana, Thin, Lena, Tkachev, Alexander, Tolmanis, Ivars, Tsarynna, Nataliia, Tsonev, Nikolay, Tulassay, Zsolt, Ueo, Tetsuya, Unal, Nalan, Valuyskikh, Ekaterina, Van Dop, Willemijn, Vasilevskaya, Olga, Venugopal, Kannan, Verstockt, Bram, Viamonte, Manuel, Vyshyvanyuk, Vira, Wang, Bangmao, Wang, Xiaoyan, Wang, Jiangbin, Wang, Yufang, Wei, Shu-chen, Weisshof, Roni, Wisam, Sbeit, Wojcik, Katarzyna, Yakovlev, Alexey, Yasuda, Hiroshi, Ye, Byongduk, Yen, Hsu-heng, Yilmaz, Hasan, Yokoyama, Yoko, Yoon, Hyuk, Yoshida, Takeshi, Yurkiv, Andriy, Zaha, Osamu, Zaltman, Cyrla, Zhan, Qiang, Zhang, Hongjie, Zhang, Shutian, Zhu, Lanxiang, Zou, Duowu, Zou, Xiaoping, Zureikat, Firas, Rubin, David T, Allegretti, Jessica R, Panés, Julián, Shipitofsky, Nicole, Yarandi, Shadi S, Huang, Kuan-Hsiang Gary, Germinaro, Matthew, Wilson, Rebbecca, Zhang, Hongyan, Johanns, Jewel, Feagan, Brian G, Hisamatsu, Tadakazu, Lichtenstein, Gary R, Bressler, Brian, Peyrin-Biroulet, Laurent, Sands, Bruce E, and Dignass, Axel

Published
2025
Full Text
View/download PDF

41. Case report: Fatal long-term intoxication by 2,4-dinitrophenol and anabolic steroids in a young bodybuilder with muscle dysmorphia

Author

Coralie Hermetet, Marine Jourdan, Alain Baert, Laurie Gheddar, Alice Ameline, Pascal Kintz, and Renaud Bouvet

Subjects
2,4-dinitrophenol, muscle dysmorphia, bodybuilding, chronic poisoning, aggressive resuscitation therapeutics, case report, Public aspects of medicine, RA1-1270
Abstract

A case of chronic intoxication by 2,4-dinitrophenol (2,4-DNP) is reported in a 21-year-old bodybuilder, also known as an abuser of anabolic steroids, who died after ingesting 2 grams of this substance after 6 months of repeated consumption. The bodybuilder presented the triad of symptoms – tachycardia, tachypnoea, profuse sweating – from 6 months before his death, and was hospitalised for multiple organ failure 4 months before his death. Medical staff attributed this serious episode to his consumption of 2,4-DNP. Although the triad of symptoms persisted, he denied continuing to consume 2,4-DNP during consultations with his general practitioner, who therefore looked into a possible diagnosis of endocrine or tumour disorder. The bodybuilder died of multi-organ failure. The autopsy found diffuse visceral congestion and yellowish colouration of integuments. Toxicology demonstrated not only lethal acute 2,4-DNP intoxication (blood concentration was 88 mg/L), but also chronic intoxication (segmental hair concentrations were 5.1 to 25.5 ng/mg). Different anabolic steroids were also identified in the hair. Continued uncontrolled consumption of 2,4-DNP despite the consequences for his health, combined with an obvious preoccupation with his physical appearance, supported the suspected diagnosis of “muscle dysmorphia”, a psychiatric disorder in which dangerous substances are trivialised. Primary care professionals, the first to come into contact with intoxication cases, should receive training on how to detect and manage cases with symptomology that mimics 2,4-DNP use. A large study evaluating the outcomes of acute intoxication cases if “aggressive” management had been immediately implemented may help us determine the optimal course of action that minimises fatalities.

Published
2024
Full Text
View/download PDF

42. A genome-wide gene-environment interaction study of breast cancer risk for women of European ancestry

Author

Middha, Pooja, Wang, Xiaoliang, Behrens, Sabine, Bolla, Manjeet K, Wang, Qin, Dennis, Joe, Michailidou, Kyriaki, Ahearn, Thomas U, Andrulis, Irene L, Anton-Culver, Hoda, Arndt, Volker, Aronson, Kristan J, Auer, Paul L, Augustinsson, Annelie, Baert, Thaïs, Freeman, Laura E Beane, Becher, Heiko, Beckmann, Matthias W, Benitez, Javier, Bojesen, Stig E, Brauch, Hiltrud, Brenner, Hermann, Brooks-Wilson, Angela, Campa, Daniele, Canzian, Federico, Carracedo, Angel, Castelao, Jose E, Chanock, Stephen J, Chenevix-Trench, Georgia, Cordina-Duverger, Emilie, Couch, Fergus J, Cox, Angela, Cross, Simon S, Czene, Kamila, Dossus, Laure, Dugué, Pierre-Antoine, Eliassen, A Heather, Eriksson, Mikael, Evans, D Gareth, Fasching, Peter A, Figueroa, Jonine D, Fletcher, Olivia, Flyger, Henrik, Gabrielson, Marike, Gago-Dominguez, Manuela, Giles, Graham G, González-Neira, Anna, Grassmann, Felix, Grundy, Anne, Guénel, Pascal, Haiman, Christopher A, Håkansson, Niclas, Hall, Per, Hamann, Ute, Hankinson, Susan E, Harkness, Elaine F, Holleczek, Bernd, Hoppe, Reiner, Hopper, John L, Houlston, Richard S, Howell, Anthony, Hunter, David J, Ingvar, Christian, Isaksson, Karolin, Jernström, Helena, John, Esther M, Jones, Michael E, Kaaks, Rudolf, Keeman, Renske, Kitahara, Cari M, Ko, Yon-Dschun, Koutros, Stella, Kurian, Allison W, Lacey, James V, Lambrechts, Diether, Larson, Nicole L, Larsson, Susanna, Le Marchand, Loic, Lejbkowicz, Flavio, Li, Shuai, Linet, Martha, Lissowska, Jolanta, Martinez, Maria Elena, Maurer, Tabea, Mulligan, Anna Marie, Mulot, Claire, Murphy, Rachel A, Newman, William G, Nielsen, Sune F, Nordestgaard, Børge G, Norman, Aaron, O’Brien, Katie M, Olson, Janet E, Patel, Alpa V, Prentice, Ross, Rees-Punia, Erika, Rennert, Gad, Rhenius, Valerie, Ruddy, Kathryn J, and Sandler, Dale P

Subjects
Biomedical and Clinical Sciences, Oncology and Carcinogenesis, Breast Cancer, Cancer, Estrogen, Prevention, Women's Health, Human Genome, Genetics, Aging, Cancer Genomics, 2.1 Biological and endogenous factors, Adult, Female, Humans, Gene-Environment Interaction, Genetic Predisposition to Disease, Breast Neoplasms, Bayes Theorem, Genome-Wide Association Study, Risk Factors, Polymorphism, Single Nucleotide, Case-Control Studies, Breast cancer, Gene-environment interactions, Genetic epidemiology, European ancestry, CTS Consortium, ABCTB Investigators, kConFab Investigators, Oncology & Carcinogenesis, Oncology and carcinogenesis
Abstract

BackgroundGenome-wide studies of gene-environment interactions (G×E) may identify variants associated with disease risk in conjunction with lifestyle/environmental exposures. We conducted a genome-wide G×E analysis of ~ 7.6 million common variants and seven lifestyle/environmental risk factors for breast cancer risk overall and for estrogen receptor positive (ER +) breast cancer.MethodsAnalyses were conducted using 72,285 breast cancer cases and 80,354 controls of European ancestry from the Breast Cancer Association Consortium. Gene-environment interactions were evaluated using standard unconditional logistic regression models and likelihood ratio tests for breast cancer risk overall and for ER + breast cancer. Bayesian False Discovery Probability was employed to assess the noteworthiness of each SNP-risk factor pairs.ResultsAssuming a 1 × 10-5 prior probability of a true association for each SNP-risk factor pairs and a Bayesian False Discovery Probability

Published
2023

43. Aggregation tests identify new gene associations with breast cancer in populations with diverse ancestry

Author

Mueller, Stefanie H, Lai, Alvina G, Valkovskaya, Maria, Michailidou, Kyriaki, Bolla, Manjeet K, Wang, Qin, Dennis, Joe, Lush, Michael, Abu-Ful, Zomoruda, Ahearn, Thomas U, Andrulis, Irene L, Anton-Culver, Hoda, Antonenkova, Natalia N, Arndt, Volker, Aronson, Kristan J, Augustinsson, Annelie, Baert, Thais, Freeman, Laura E Beane, Beckmann, Matthias W, Behrens, Sabine, Benitez, Javier, Bermisheva, Marina, Blomqvist, Carl, Bogdanova, Natalia V, Bojesen, Stig E, Bonanni, Bernardo, Brenner, Hermann, Brucker, Sara Y, Buys, Saundra S, Castelao, Jose E, Chan, Tsun L, Chang-Claude, Jenny, Chanock, Stephen J, Choi, Ji-Yeob, Chung, Wendy K, Colonna, Sarah V, Cornelissen, Sten, Couch, Fergus J, Czene, Kamila, Daly, Mary B, Devilee, Peter, Dörk, Thilo, Dossus, Laure, Dwek, Miriam, Eccles, Diana M, Ekici, Arif B, Eliassen, A Heather, Engel, Christoph, Evans, D Gareth, Fasching, Peter A, Fletcher, Olivia, Flyger, Henrik, Gago-Dominguez, Manuela, Gao, Yu-Tang, García-Closas, Montserrat, García-Sáenz, José A, Genkinger, Jeanine, Gentry-Maharaj, Aleksandra, Grassmann, Felix, Guénel, Pascal, Gündert, Melanie, Haeberle, Lothar, Hahnen, Eric, Haiman, Christopher A, Håkansson, Niclas, Hall, Per, Harkness, Elaine F, Harrington, Patricia A, Hartikainen, Jaana M, Hartman, Mikael, Hein, Alexander, Ho, Weang-Kee, Hooning, Maartje J, Hoppe, Reiner, Hopper, John L, Houlston, Richard S, Howell, Anthony, Hunter, David J, Huo, Dezheng, Ito, Hidemi, Iwasaki, Motoki, Jakubowska, Anna, Janni, Wolfgang, John, Esther M, Jones, Michael E, Jung, Audrey, Kaaks, Rudolf, Kang, Daehee, Khusnutdinova, Elza K, Kim, Sung-Won, Kitahara, Cari M, Koutros, Stella, Kraft, Peter, Kristensen, Vessela N, Kubelka-Sabit, Katerina, Kurian, Allison W, Kwong, Ava, Lacey, James V, Lambrechts, Diether, and Le Marchand, Loic

Subjects
Biological Sciences, Genetics, Biotechnology, Clinical Research, Women's Health, Cancer, Breast Cancer, 2.1 Biological and endogenous factors, 4.1 Discovery and preclinical testing of markers and technologies, Humans, Female, Breast Neoplasms, Genetic Predisposition to Disease, Black People, Genetic Testing, Genome-Wide Association Study, Polymorphism, Single Nucleotide, Formins, Breast cancer susceptibility, Diverse ancestry, Rare variants, Gene regulation, Genome-wide association study, NBCS Collaborators, CTS Consortium, ABCTB Investigators, Clinical Sciences
Abstract

BackgroundLow-frequency variants play an important role in breast cancer (BC) susceptibility. Gene-based methods can increase power by combining multiple variants in the same gene and help identify target genes.MethodsWe evaluated the potential of gene-based aggregation in the Breast Cancer Association Consortium cohorts including 83,471 cases and 59,199 controls. Low-frequency variants were aggregated for individual genes' coding and regulatory regions. Association results in European ancestry samples were compared to single-marker association results in the same cohort. Gene-based associations were also combined in meta-analysis across individuals with European, Asian, African, and Latin American and Hispanic ancestry.ResultsIn European ancestry samples, 14 genes were significantly associated (q

Published
2023

47. FaceTuneGAN: Face Autoencoder for Convolutional Expression Transfer Using Neural Generative Adversarial Networks

Author

Olivier, Nicolas, Baert, Kelian, Danieau, Fabien, Multon, Franck, and Avril, Quentin

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Graphics
Abstract

In this paper, we present FaceTuneGAN, a new 3D face model representation decomposing and encoding separately facial identity and facial expression. We propose a first adaptation of image-to-image translation networks, that have successfully been used in the 2D domain, to 3D face geometry. Leveraging recently released large face scan databases, a neural network has been trained to decouple factors of variations with a better knowledge of the face, enabling facial expressions transfer and neutralization of expressive faces. Specifically, we design an adversarial architecture adapting the base architecture of FUNIT and using SpiralNet++ for our convolutional and sampling operations. Using two publicly available datasets (FaceScape and CoMA), FaceTuneGAN has a better identity decomposition and face neutralization than state-of-the-art techniques. It also outperforms classical deformation transfer approach by predicting blendshapes closer to ground-truth data and with less of undesired artifacts due to too different facial morphologies between source and target.

Published
2021

Catalog

Books, media, physical & digital resources
See catalog results