# Comparing Completion Time, Accuracy, and Satisfaction in Virtual Reality vs. Desktop Implementation of the Common Coordinate Framework Registration User Interface (CCF RUI)

**Authors:** Andreas Bueckle, Kilian Buehling, Patrick C. Shih, Katy Börner  
**arXiv:** [2102.12030](https://arxiv.org/abs/2102.12030)  
**Date:** 2021-02-24  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2102.12030v1/

## Key points

- The Registration User Interface (RUI) was developed for stakeholders in the Human Biomolecular Atlas Program (HuBMAP) to register tissue blocks by size, position, and orientation.
- RUI has been used by tissue mapping centers across the HuBMAP consortium to register a total of 45 kidney, spleen, and colon tissue blocks.
- A user study involving 42 human subjects was conducted to compare the performance of three setups: traditional 2D Desktop implementation, VR Tabletop, and VR Standup.
- Results showed that while VR Tabletop and VR Standup users were about three times as fast and about a third more accurate in terms of rotation than 2D Desktop users, there were no significant differences for position accuracy.
- The 2D Desktop version is well-suited for registering tissue blocks at a speed and accuracy that meets the needs of experts performing tissue dissection. It is also cheaper, easier to learn, and more practical for wet-bench environments than the VR setups.
- Future work includes testing interventions to help users improve their performance during the experiment by running a study with a “reflective” phase where the user sees a visualization of their own performance data from previous tasks before completing a second set of tasks.

## Summary

The Registration User Interface (RUI) was developed to enable stakeholders in the Human Biomolecular Atlas Program (HuBMAP) to register tissue blocks by size, position, and orientation. The RUI has been used by tissue mapping centers across the HuBMAP consortium to register a total of 45 kidney, spleen, and colon tissue blocks. In this paper, the authors compare three setups for registering one 3D tissue block object to another 3D reference organ (target) object. The first setup is a traditional 2D Desktop implementation featuring a screen, mouse, and keyboard interface. The remaining setups are both virtual reality (VR) versions of the RUI: VR Tabletop, where users sit at a physical desk; VR Standup, where users stand upright. A user study involving 42 human subjects completing increasingly difficult tasks was conducted to compare the performance of these three setups. Results showed that while VR Tabletop and VR Standup users were about three times as fast and about a third more accurate in terms of rotation than 2D Desktop users, there were no significant differences for position accuracy. The performance values for the 2D Desktop version confirm that it is well-suited for registering tissue blocks at a speed and accuracy that meets the needs of experts performing tissue dissection. Additionally, the 2D Desktop setup is cheaper, easier to learn, and more practical for wet-bench environments than the VR setups. The paper also presents an overview of materials and methods used in conducting this study including task difficulty metrics generation using synthetic tasks generation; human performance metrics such as plateau; satisfaction score computation; power analysis conducted prior to running the experiment to determine the number of subjects required to achieve significant results; typical user study design featuring pre- and post-questionnaires with tutorial and experiment tasks; all subjects using Alienware laptop running Unity game engine on Windows 10 with secondary monitor attached for ease of configuring individual steps of the experiment; and data collection using a custom C# script writing data to a CSV file at a frequency of 10 Hz every time the user pressed a button. Future work includes testing how interventions could help users improve their performance during the experiment by running a study with a “reflective” phase where the user sees a visualization of their own performance data from previous tasks before completing a second set of tasks. The authors aim to use interactive data visualizations to help users formulate strategies to improve their performance in terms of position accuracy, rotation accuracy, and completion time. Additionally, given the detailed telemetry data collected from RUI users (especially those in VR), an intervention where users can see their own movement as well as the position and rotation of the tissue and target blocks over time is planned. The HuBMAP project aims to create an unprecedented reference atlas of healthy human bodies at single-cell resolution which will capture spatial information about cells and tissues in detail so as develop novel approaches towards curing diseases through hundreds researchers from several dozen institutions across U.S., Europe & other countries worldwide .