This paper presents a Task-Free eye-tracking dataset for Dynamic Point Clouds (TF-DPC) aimed at investigating visual attention. The is composed of eye gaze and head movements collected from 24 participants observing 19 scanned dynamic point clouds in Virtual Reality (VR) environment with 6 degrees freedom. We compare the saliency maps generated this those prior task-dependent experiment (focused on quality assessment) to explore how high-level tasks influence human To measure similarity...

This article focuses on motion prediction for point cloud sequences in the challenging case of deformable 3D objects, such as human body motion. First, we investigate challenges caused by shapes and complex motions present this type representation, with ultimate goal understanding technical limitations state-of-the-art models. From understanding, propose an improved architecture objects. Specifically, to handle shapes, a graph-based approach that learns exploits spatial structure clouds...

Virtual reality (VR) applications target high-quality and zero-latency scene navigation to provide users with a full-immersion sensation within scene. From network perspective, this requires transmission of the omnidirectional content in its entirety, at high resolution, which is not always feasible bandwidth-limited networks. In work, we propose an optimal strategy for virtual able fulfill bandwidth requirements, while optimizing end-user quality experienced navigation. further detail,...

In Virtual Reality (VR) applications, understanding how users explore the omnidirectional content is important to optimize creation, develop user-centric services, or even detect disorders in medical applications. Clustering based on their common navigation patterns a first direction understand behavior. However, classical clustering techniques fail identifying this paths, since they are usually focused minimizing simple distance metric. paper, we argue that metric does not necessarily...

This work presents an explorative behavioural analysis of users navigating in immersive space aimed at enabling the next-generation multimedia systems. Our main goal is to understand how user experience content with 6-Degrees-of-Freedom (DoF) affected not only by visual and its quality but also disposition user. We based our investigations on traditional statistical metrics, techniques that have been already used for 6-DoF, as well adapted 3-DoF tools be this new context. show limitation...

In recent years, a large variety of online communication tools have emerged, including social Virtual Reality (VR) platforms for interacting in virtual world with participants being represented as avatars. Given their popularity, an active area research focuses on improving the user experience these experiences. To enable experimentation at scale platforms, it is however essential to collect behavioural data (e.g. movements and audio information). this work, we present toolchain that enables...

With the overarching goal of developing user-centric Virtual Reality (VR) systems, a new wave studies focused on understanding how users interact in VR environments has recently emerged. Despite intense efforts, however, current literature still does not provide right framework to fully interpret and predict users’ trajectories while navigating scenes. This work advances state-of-the-art both study behaviour system design. In more detail, we complement datasets by presenting publicly...

Social Virtual Reality (VR) applications represent a big step forward in the field of remote communication. VR provides possibility for participants to explore and interact with virtual environments objects, feelings full sense immersion, being together. Understanding how user behaviour is influenced by shared space its elements becomes key design optimize novel immersive experiences. This paper presents behavioural analysis navigating 6 degrees freedom social movie. Specifically, we analyse...

In this paper, we propose an end-to-end learning network to predict future frames in a point cloud sequence. As main novelty, initial layer learns topological information of clouds as geometric features, form representative spatiotemporal neighborhoods. This module is followed by multiple Graph-RNN cells. Each cell dynamics (i.e., RNN states) processing each jointly with its neighbours. We tested the performance MNIST dataset moving digits, synthetic human bodies motions, and JPEG dynamic...

The latest social VR technologies have enabled users to attend traditional media and arts performances together while being geographically removed, making such experiences accessible despite budget, distance, other restrictions. In this work, we aim at improving the way remote are shared by designing evaluating a theatre lobby which serves as space for gather, interact, relive common experience of watching virtual opera. We conducted an initial test with experts ($\mathrm{N}=10$, i.e.,...

To cope with the large bandwidth and low-latency requirements, Virtual Reality (VR) systems are steering toward user-centric in which coding, streaming, possibly rendering personalized to final user. The success of these VR mainly relies on ability anticipate viewers navigation. This has motivated a attention studying prediction user's movements experience. However, most work lack proper exhaustive behavioural analysis scenario, leaving many key-behavioural questions unsolved unexplored: Can...

Immersive reality technologies, such as Virtual and Augmented Reality, have ushered a new era of user-centric systems, in which every aspect the coding--delivery--rendering chain is tailored to interaction users. Understanding actual interactivity behaviour users still an open challenge key step enabling system. Our main goal extend applicability existing behavioural methodologies for studying user navigation case 6 Degree-of-Freedom (DoF). Specifically, we first compare 6-DoF with its 3-DoF...

The general understanding of user behaviour has been often overlooked in the field Virtual Reality (VR) and Extended (XR) at large. In this work, we want to fill gap by exploring relationship between way which users navigate immersive content predictability their trajectories. Inspired works from social science, our key assumption is that there are navigation trajectories can be accurately predicted, while others exhibit eclectic patterns more challenging anticipate. However, it not yet...

Thanks to recent advances in computer graphics, wearable technology, and connectivity, Virtual Reality (VR) has landed our daily life. A key novelty VR is the role of user, which turned from merely passive entirely active. Thus, improving any aspect coding-delivery-rendering chain starts with need for understanding user behaviour. To do so, we investigate navigation trajectories users within a 6-Degrees-of-Freedom (DoF) environment. Specifically, main differences similarities between 3 6-DoF...

The advent in our daily life of Extended Reality (XR) technologies, such as Virtual and Augmented Reality, has led to the rise user-centric systems, offering higher level interaction presence virtual environments. In this context, understanding actual interactivity users is still an open challenge a key step enabling system. work, goal construct efficient clustering tool for 6 Degree-of-Freedom (DoF) navigation trajectories by extending applicability existing behavioural tool. Specifically,...

The explainability of a robot's actions is crucial to its acceptance in social spaces. Explaining why robot fails complete given task particularly important for non-expert users be aware the capabilities and limitations. So far, research on explaining failures has only considered generating textual explanations, even though several studies have shown benefits multimodal ones. However, simple combination multiple modalities may lead semantic incoherence between information across different -...

Immersive reality technologies, such as Virtual and Augmented Reality, have ushered a new era of user-centric systems, in which every aspect the coding--delivery--rendering chain is tailored to interaction users. Understanding actual interactivity behaviour users still an open challenge key step enabling system. Our main goal extend applicability existing behavioural methodologies for studying user navigation case 6 Degree-of-Freedom (DoF). Specifically, we first compare 6-DoF with its 3-DoF...

This paper aims at bringing some light and understanding to the field of deep learning for dynamic point cloud processing. Specifically, we focus on hierarchical features aspect, with ultimate goal which are learned different stages process what their meaning is. Last, bring clarity how components network affect importance a successful model. study is conducted prediction tasks, useful predicting coding applications.