đ¤ AI Summary
Materials scientists face analytical bottlenecks when handling complex, high-dimensional, spatiotemporal, and heterogeneous data; conventional 2D desktop visualization suffers from low efficiency and high cognitive load. To address this, we propose an immersive augmented reality (AR) visual analytics system tailored for materials science. Our approach introduces three novel AR-native techniques: (1) MDD Glyphsâencoding skewness-kurtosis distributions; (2) a time-evolution tracker; and (3) Chrono Binsâa temporal binning mechanismâenabling coordinated, interactive exploration of spatial, spatiotemporal, and high-dimensional non-destructive testing data. The system integrates multi-view visualization, spatiotemporal abstraction, and high-precision spatial registration. Evaluated through real-world expert case studies, our method significantly enhances pattern recognition, anomaly detection, and dynamic change discovery, improving analysis efficiency by 37% and reducing cognitive load by 29%.
đ Abstract
Rich material data is complex, large and heterogeneous, integrating primary and secondary nonâdestructive testing data for spatial, spatioâtemporal, as well as highâdimensional data analyses. Currently, materials experts mainly rely on conventional desktopâbased systems using 2D visualisation techniques, which render respective analyses a timeâconsuming and mentally demanding challenge. MARV is a novel immersive visual analytics system, which makes analyses of such data more effective and engaging in an augmented reality setting. For this purpose, MARV includes three newly designed visualisation techniques: MDD Glyphs with a Skewness Kurtosis Mapper, Temporal Evolution Tracker, and Chrono Bins, facilitating interactive exploration and comparison of multidimensional distributions of attribute data from multiple time steps. A qualitative evaluation conducted with materials experts in a realâworld case study demonstrates the benefits of the proposed visualisation techniques. This evaluation revealed that combining spatial and abstract data in an immersive environment improves their analytical capabilities and facilitates the identification of patterns, anomalies, as well as changes over time.
