← All special sessions

Deviceless Multidimensional Human–Object Perception: Toward Device-Free Sensing and Interaction

Abstract

This special session explores emerging paradigms in device-free human–object perception, focusing on the extraction of multidimensional attributes, including spatial location, motion dynamics, shape and size, and interaction semantics without requiring targets to carry any physical device. By leveraging vision-based techniques and emerging non-vision wireless modalities such as RGB, RGB-D camera, mmWave radar, LiDAR, WiFi Channel State Information (CSI), and 6G Integrated Sensing and Communication (ISAC) base station, the workshop aims to advance HCI community’s goal of creating human-centred technologies that organically augment how we live, work, and learn.

Aligned with the vision of the IHCI, this session brings together interdisciplinary perspectives spanning wireless sensing, computer vision, machine learning, and human-centred design with emphasis on seamless HCI.

The session will feature technical presentations and a panel discussion addressing key challenges such as interpretability, latency, environmental robustness, ethical considerations and so on. By bridging sensing technologies with human-computer interaction principles, the session aims to catalyse next-generation interfaces that are pervasive, ubiquitous, intuitive, privacy and context-aware.

Rationale and Scope

The shift toward device-free sensing represents a fundamental transformation in human–computer interaction. Conventional systems rely heavily on wearable or handheld devices, introducing limitations related to usability, maintenance, energy constraints, and user compliance. However, the future of seamless interaction lies in ambient intelligence, where the environment itself perceives and responds to user needs organically. Emerging sensing paradigms, particularly in wireless and RF domains enable such passive perception of human activity and object characteristics.

This workshop directly addresses IHCI’s goal of integrating technology with human-centred design by enabling invisible interfaces that perceive and respond to human presence and intent without explicit user instrumentation. This includes recent advances in wireless sensing such as mmWave radar sensing (micro-Doppler, range-angle imaging), WiFi CSI-based activity recognition, LiDAR-based 3D scene understanding and 6G ISAC sensing frameworks. Advances in these frameworks enable multidimensional perception beyond simple activity recognition, including spatial-temporal behaviour modelling and object interaction understanding. In addition, non-vision modalities not only solve occlusion and low-light issues but also fundamentally address the growing privacy concerns associated with pervasive camera use in living spaces. However, challenges remain in real-time, low-latency inference, cross-domain generalization, multimodal fusion, and human-centric interpretability. This session is timely in consolidating these rapidly evolving research directions towards one HCI perspective.

Intended Audience and Anticipated Impact

This session targets a broad interdisciplinary audience, including academic researchers, HCI practitioners, UX/UI designers, IoT engineers, telecommunications experts, and industry practitioners in smart healthcare, mobility, and ambient intelligence, designers and social scientists interested in human-centred and ethical AI. We anticipate at least 30 high-quality paper submissions, of which 4–6 will be accepted. We also aim to secure 1 keynote or invited talk from a leading expert in this field.

This session will foster cross-disciplinary collaboration among RF engineers, AI researchers, and HCI designers, establishing a unified research agenda for device-free multidimensional perception and influencing future research in 6G sensing, ambient intelligence, and invisible interfaces. Building on a shared foundation, the anticipated impact includes accelerating the development of privacy-centric smart healthcare monitoring, seamless smart home automation, and safer industrial IoT environments, ultimately setting new standards for the ethical and effective use of passive sensing in HCI.

The chairs are committed to ensuring a diverse and inclusive environment. We will actively reach out to underrepresented groups in the fields of HCI to solicit submissions.

Specialized Topics

Prospective authors are invited to submit original research, methodological papers, and practical case studies. Topics of interest include, but are not limited to:

Core Perception, Sensing and Modality

  • Device-free human activity recognition (HAR)
  • Object detection and classification
  • Multidimensional sensing: location, velocity, size, pose, and interaction semantics
  • Micro-Doppler and gesture/activity signature analysis
  • Using single or fusion of multiple modalities such as camera-based systems, WiFi, mmWave radar, LiDAR and 6G ISAC to perform the tasks above

Machine Learning and Representation

  • Domain adaptation and transfer learning for sensing
  • Self-supervised and contrastive learning for various modalities
  • Lightweight and real-time inference models
  • Representation learning for multidimensional sensing tensors

Systems and Deployment

  • Real-time edge AI for sensing systems
  • Embedded and low-power sensing architectures
  • Large-scale deployment in smart homes, cities, and healthcare

Human-Centred and Ethical Considerations

  • Privacy-preserving sensing techniques
  • Explainability and interpretability in sensing models
  • User trust, transparency, and acceptance
  • Accessibility and inclusive interaction design

Chairs

  • Dr. David Chieng, University of Nottingham Ningbo China
  • Dr. Yi Zhong, Beijing Institute of Technology, China
  • Professor Alan Marshall, University of Liverpool, U.K.
  • Professor Carlos J. Bernardos, Universidad Carlos III de Madrid, Spain
  • Dr. Chee Keong Tan, Monash University, Malaysia Campus, Malaysia