Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Bruno Dumas has a total of 51 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Atelier à la conférence IHM 2025 : RA Permanente

    by Maxime Cauz, Thibaut Septon, Elise Hallaert, Theo Leclercq, Bruno Dumas, Charles Bailly, Clement Tyminski, Matias Peraza, Sophie Lepreux, Emmanuel Dubois

    As we move towards more ubiquitous computing, the concept of pervasive augmented reality (PAR) could lead to a major evolution in the relationship between humans, computing and the world. The experience of a continuously augmented world can have both benefits and undesirable consequences for users' lives, and raises many questions in multiple areas. In this workshop, we wanted to bring together all IHM'25 conference participants who are concerned or enthusiastic about discussing this topic. The aim was to draw on collective intelligence to identify the interdisciplinary challenges that remain to be resolved in order to enable the implementation of these technologies in everyday life, but also to define the necessary safeguards. Is PAR too techno-enthusiastic? All of these elements were grouped into categories to define a set of future major areas of research around permanent augmented reality. This document is in French as the conference is a French-speaking international conference.

  • Open Access

    Computer Assisted Diagnosis: does it make sense for Lungs Tumor Detection? A Qualitative study with Pulmonologists.

    by Robin Ghyselinck, Thibaut Septon, Nicolas Geuens, Margaux Mannaerts, Julie Vijghen, Samy Lachkar, Benoit Frenay, Bruno Dumas

  • Open Access

    TapStrapGest: Elicitation and Recognition of Ring-based Multi-Finger Gestures

    by Guillem Cornella-Barba, Bruno Dumas, Mehdi Ousmer, Santiago Villarreal-Narvaez, Jean Vanderdonckt, Eudald Sangenis, Adrien Chaffangeon Caillet

    We introduce TapStrapGest , a novel solution for customizable ring-based multi-finger gestures, encompassing the process from gesture elicitation to gesture recognition. Recognizing the growing demand for intuitive and customizable gesture interaction with fingers, TapStrapGest uses Tap Strap to enable users to perform simple and complex multi-finger gestures using smart rings. We conducted a gesture elicitation study, detailing the systematic process of soliciting and refining a custom set of user-defined ring-based finger gestures through participatory design and ergonomic considerations, including thinking time, goodness of fit, and memorization. Subsequently, we delve into the technical underpinnings of gesture recognition. We reduce the dimensionality of a dataset of 27 gesture classes from 21 to 15 by filtering, then from 15 to 5 by a Principal Component Analysis. We implement and compare four machine learning algorithms to show that a Quadratic Discriminant Analysis (precision=99.33%, recall=99.26%, and F1-score=99.26%) outperforms three other machine learning classifiers, i.e., a Linear Discriminant Analysis, a Support Vector Machines, and a Random Forest, as well as existing recognizers from the literature, to accurately recognize such gestures without the need to call for Deep Learning. Through a performance analysis, we demonstrate that TapStrapGest is a versatile and admissible solution for ring-based multi-finger gesture interaction, opening avenues for "eyes-free" or "screen-free" human-computer interaction in various domains.


  • Open Access

    On the effectiveness of Rotation-Equivariance in U-Net: A Benchmark for Image Segmentation

    by Robin Ghyselinck, Valentin Delchevalerie, Bruno Dumas, Benoît Frénay

    Numerous studies have recently focused on incorporating different variations of equivariance in Convolutional Neural Networks (CNNs). In particular, rotation-equivariance has gathered significant attention due to its relevance in many applications related to medical imaging, microscopic imaging, satellite imaging, industrial tasks, etc. While prior research has primarily focused on enhancing classification tasks with rotation equivariant CNNs, their impact on more complex architectures, such as U-Net for image segmentation, remains scarcely explored. Indeed, previous work interested in integrating rotation-equivariance into U-Net architecture have focused on solving specific applications with a limited scope. In contrast, this paper aims to provide a more exhaustive evaluation of rotation equivariant U-Net for image segmentation across a broader range of tasks. We benchmark their effectiveness against standard U-Net architectures, assessing improvements in terms of performance and sustainability (i.e., computational cost). Our evaluation focuses on datasets whose orientation of objects of interest is arbitrary in the image (e.g., Kvasir-SEG), but also on more standard segmentation datasets (such as COCO-Stuff) as to explore the wider applicability of rotation equivariance beyond tasks undoubtedly concerned by rotation equivariance. The main contribution of this work is to provide insights into the trade-offs and advantages of integrating rotation equivariance for segmentation tasks.


  • Open Access

    Exploiting Semantic Search and Object-Oriented Programming to Ease Multimodal Interface Development

    by Thibaut Septon, Santiago Villarreal-Narvaez, Xavier Devroey, Bruno Dumas

    Multimodal interaction has been adopted across various platforms and devices, with supporting tools enhancing the developer experience in developing (). While traditionally, these tools faced challenges balancing expressiveness and usability, recent progress in tends to mitigate this rule. However, adding multimodal interaction still remains challenging, especially when integrating the voice modality, and s remain to be better integrated into today’s applications. To address these challenges, we introduce a Unity tool-based system named. allows developers to use their knowledge in Object-Oriented Programming to handle the expert knowledge required to create a while allowing end users great flexibility in the natural language they can use to interact. Our contributions are: 1) the presentation of ’s architecture and its inherent concepts along with its open-source implementation, and 2) a successful evaluation of its usability for describing s through the questionnaire with twelve participants.

  • Open Access

    Engineering User Interfaces for Tailored and Monitored Movement Rehabilitation Programs

    by Corentin Reuther, Gerry Longfils, Rosny Anderson Fodja Kouanga, Maxime André, Bruno Dumas

    As a commercially available gesture tracking device, the Microsoft Kinect was already used for supporting rehabilitation programs in the field of physiotherapy. However, physiotherapists found that the device and applications on offer did not allow them to tailor and monitor exercises accurately. They point out that every patient is different and requires adapted exercises with an expert eye. This paper proposes an approach to revitalize the use of this device in physiotherapy by addressing these needs. Working closely with a physiotherapist, we developed Motion Recovery, a solution that offers precise tailoring and monitoring capabilities for physiotherapy rehabilitation programs. Scoring 82.8 on the System Usability Scale, a test providing an overview of interactive system usability, our solution demonstrated strong user acceptance among 15 patients, culminating in a qualitative evaluation with the physiotherapist.


  • Open Access

    Text readability in augmented reality: a multivocal literature review

    by Maxime Cauz, Antoine Clarinval, Bruno Dumas

    Augmented reality (AR) is making its way into many sectors. Its rapid evolution in recent years has led to the development of prototypes demonstrating its effectiveness. However, to be able to push these prototypes to the scale of fully usable applications, it is important to ensure the readability of the texts they include. To this end, we conducted a multivocal literature review (MLR) to determine the text parameters a designer can tune, as well as the contextual constraints they need to pay attention to, in relation to Optical See-Through (OST) and Video See-Through (VST) displays. We also included guidelines from device manufacturing and game engines sites to compare the current state of research in the academic and industrial worlds. The results show that parameters pertaining more to letter legibility have been extensively studied (e.g., color and size), while those pertaining to the whole text still require further research (e.g., alignment or space between lines). The former group of parameters, and their associated constraints, were assembled in the form of two decision trees to facilitate implementation of AR applications. Finally, we also concluded that there was a lack of alignment between academic and industrial recommendations.