Brain tumours differ significantly in shape, size, location, and contrast imperfections. Reliable segmentation is crucial for accurate identification and effective treatment. Recent advances in AI-based 3D medical image segmentation, such as SAM-Med3D, have improved automation; however, expert supervision remains vital for brain imaging. We present a virtual reality probe that integrates an expert-in-the-loop approach with SAM-Med3D, allowing users to iteratively enhance brain tumour segmentation by selecting points for AI refinement.

The full paper preprint can be found here: PrePrint.

Together with industry experts, we are exploring the potential of head-mounted augmented reality to facilitate safety inspections on high-rise construction sites. A particular concern in the industry is inspecting perimeter safety screens on higher levels of construction sites, intended to prevent falls of people and objects. We aim to support workers performing this inspection task by tracking which parts of the safety screens have been inspected. We use machine learning to automatically detect gaps in the perimeter screens that require closer inspection and remediation and to automate reporting. This work-in-progress paper describes the problem, our early progress, concerns around worker privacy, and the possibilities to mitigate these.

The full paper preprint can be found here: PrePrint.

Augmented Reality (AR) is touted to be beneficial in supporting situated information display, allowing virtual information panels to be overlaid on real-world scenes. People must then use their spatial memory to navigate among these virtual panels effectively. While spatial memory has been studied in physical environments (wall displays) and virtual reality environments, there has been little research on how physical surroundings might affect memorisation of virtual content in a mixed environment like AR. Therefore, we provide the first AR study of spatial memory, comparing two different room settings with two different situated layouts of virtual targets on an abstract spatial memory task. We find that participants recall spatial patterns with greater accuracy and higher subjective ratings in a room with furniture compared to an empty room. Our findings lead to important design implications for mixed-reality user interfaces, particularly in information-rich applications like situated analytics and small-multiples information visualisation.

The full paper preprint can be found here: PrePrint.

This paper presents the design and evaluation of GestureExplorer, an Immersive Analytics tool that supports the interactive exploration, classification and sensemaking with large sets of 3D temporal gesture data. GestureExplorer features 3D skeletal and trajectory visualisations of gestures combined with abstract visualisations of clustered sets of gestures. By leveraging the large immersive space afforded by a Virtual Reality interface our tool allows free navigation and control of viewing perspective for users to gain a better understanding of gestures. We explore a selection of classification methods to provide an overview of the dataset that is linked to a detailed view of the data that shows different visualisation modalities. We evaluate GestureExplorer with two user studies and collected feedback from participants with diverse visualisation and analytics backgrounds. Our results demonstrate the promising capability of GestureExplorer for providing a useful and engaging experience in exploring and analysing gesture data.

The full paper preprint can be found here: PrePrint.

Recent studies have explored how users of immersive visualisation systems arrange data representations in the space around them. Generally, these have focused on placement centred at eye-level in absolute room coordinates. However, work in HCI exploring full-body interaction has identified zones relative to the user’s body with different roles. We encapsulate the possibilities for visualisation view management into a design space (called “DataDancing”). From this design space we extrapolate a variety of view management prototypes, each demonstrating a different combination of interaction techniques and space use. The prototypes are enabled by a full-body tracking system including novel devices for torso and foot interaction. We explore four of these prototypes, encompassing standard wall and table-style interaction as well as novel foot interaction, in depth through a qualitative user study. Learning from the results, we improve the interaction techniques and propose two hybrid interfaces that demonstrate interaction possibilities of the design space.

The full paper preprint can be found here: PrePrint.

In immersive environments, positioning data visualisations around the user in a wraparound layout has been advocated as advantageous over flat arrangements more typical of traditional screens. However, other than limiting the distance users must walk, there is no clear design rationale behind this common practice, and little research on the impact of wraparound layouts on visualisation tasks. The ability to remember the spatial location of elements of visualisations within the display space is crucial to support visual analytical tasks, especially those that require users to shift their focus or perform comparisons. This ability is influenced by the user’s spatial memory but how spatial memory is affected by different display layouts remains unclear. In this paper, we perform two user studies to evaluate the effects of three layouts with varying degrees of curvature around the user (flat-wall, semicircular-wraparound, and circular-wraparound) on a visuo-spatial memory task in a virtual environment. The results show that participants are able to recall spatial patterns with greater accuracy and report more positive subjective ratings using flat than circular-wraparound layouts. While we didn’t find any significant performance differences between the flat and semicircular-wraparound layouts, participants overwhelmingly preferred the semicircular-wraparound layout suggesting it is a good compromise between the two extremes of display curvature.

The full paper preprint can be found here: PrePrint.

This paper presents GestureExplorer, which features versatile immersive visualisations to grant the user free control over their perspective, allowing them to gain a better understanding of gestures. It provides multiple data visualisation views, and interactive features to support analysis and exploration of gesture datasets. A pair of iterative user studies provides initial feedback from several participants, including experts on immersive visualisation, and demonstrates the potential of GestureExplorer for providing a useful and engaging experience for exploring gesture data. The full poster preprint can be found here: PrePrint.

With the increasing availability of head-mounted displays for virtual reality and augmented reality, we can create immersive maps in which the user is closer to the data. Embodiment is a key concept, allowing the user to act upon virtual objects in an immersive environment. Our work explores the use of embodied interaction for immersive maps. We propose four design considerations for embodied maps and embodied gesture interaction with immersive maps: object presence, consistent physics, human body skills, and direct manipulation. We present an example of an immersive flow map with a series of novel embodied gesture interactions, which adhere to the proposed design considerations. The embodied interactions allow users to directly manipulate immersive flow maps and explore origin-destination flow data in novel ways. Authors of immersive maps can use the four proposed design considerations for creating embodied gesture interactions. The discussed example interactions apply to diverse types of immersive maps and will hopefully incite others to invent more embodied interactions for immersive maps. The full paper preprint can be found here: PrePrint.

Nominated as Best Paper Award. We explore the adaptation of 2D small-multiples visualisation on flat screens to 3D immersive spaces. We use a ”shelves” metaphor for layout of small multiples and consider a design space across a number of layout and interaction dimensions. We demonstrate the applicability of a prototype system informed by this design space to data sets from different domains. We perform two user studies comparing the effect of the shelf curvature dimension from our design space on users’ ability to perform comparison and trend analysis tasks. Our results suggest that, with fewer multiples, a flat layout is more performant despite the need for participants to walk further. With an increase in the number of multiples, this performance difference disappears due to the time participants had to spend walking. In the latter case, users prefer a semi-circular layout over either a fully surrounding or a flat arrangement. The full paper preprint can be found here: PrePrint.