Secure Text Entry using a Virtual Radial Keyboard with Dynamically Resized Keys and Non-Intrusive Randomization
Yuxuan Huang, Qiao Jin, Tongyu Nie ∗, Victoria Interrante ∗, Evan Suma Rosenberg ∗ · 2026
Read Secure Text Entry using a Virtual Radial Keyboard with Dynamically Resized Keys and Non-Intrusive Randomization by Yuxuan Huang, Qiao Jin, Tongyu Nie ∗, Victoria Interrante ∗, Evan Suma Rosenberg ∗ (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-05516.txt.
Introduction
Yuxuan Huang and 4 co-authors design, in a 2026 paper, a radial virtual keyboard meant to keep VR text entry both secure and usable. Typing passwords in virtual reality risks observers inferring keystrokes from controller motion. An earlier secure method, the authors show, still had a vulnerability.
Their keyboard arranges letters in a circle, selected by rotating a controller. Keys expand dynamically to ease precise selection. After each keystroke the whole layout rotates randomly, preserving relative positions while hiding absolute ones from inference attacks. In a study with thirty users, prediction resistance improved strongly. Entry was slower and felt harder than QWERTY, though feedback suggested practice would close much of that gap.
Catalog record
| CREATOR | Yuxuan Huang, Qiao Jin, Tongyu Nie ∗, Victoria Interrante ∗, Evan Suma Rosenberg ∗ |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 61 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-05516
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Secure Text Entry using a Virtual Radial Keyboard with Dynamically Resized Keys and Non-Intrusive Randomization Yuxuan Huang, Qiao Jin, Tongyu Nie ∗, Victoria Interrante ∗, Evan Suma Rosenberg ∗ Abstract As virtual reality (VR) becomes more widely adopted, secure and efficient text entry is an increasingly critical need. In this paper, we identify a vulnerability in a state-of-the-art secure VR text entry method and introduce a novel virtual radial keyboard designed to achieve a balance between security with usability. Keys are arranged alphabetically in a circular layout, with each key selected by controller rotation and dynamically expanding to facilitate precise selection. A randomized rotation mechanism shifts the keyboard after each keystroke, preserving relative key positions while disrupting absolute spatial mappings to protect against inference attacks. We conducted a within-subject study (N=30) comparing our method with the prior secure technique and a standard QWERTY keyboard. Results showed that the radial keyboard significantly improves resistance to keystroke prediction attacks while incurring a tradeoff in entry speed and subjective workload due to the unfamiliar non-QWERTY layout. However, both quantitative trends and qualitative feedback indicate strong potential for performance improvements with practice. We also discuss design implications, possible interface refinements, and directions for future work, including layout variations and visual enhancements. \onlineid 1596 \vgtccategory Research \vgtcinsertpkg \teaser The radial keyboard proposed in this paper (left) vs. a state-of-the-art secure VR text entry method “Intermittent Starting Point Randomization” [ 43 ] (middle) vs. standard QWERTY keyboard (right). Introduction keywords Virtual Reality, Text Entry, Security, Privacy, Usability. With Virtual Reality (VR) technology becoming more accessible and widely adopted in various fields [ 18 ] , VR has demonstrated great potential in playing a more important role in our work and life. As VR applications expand beyond entertainment into areas like office work [ 14 ] , collaboration [ 22 ] , training and education [ 23 ] , the need to support essential tasks such as text entry is becoming increasingly critical [ 8 ] . However, significant challenges remain in terms of both usability and security for VR text entry [ 24 , 43 ] . Text entry in VR is inherently less efficient compared to using physical keyboards or touchpads. The root cause lies in the difficulty in precisely selecting the keys in an efficient manner as with the physical keyboard [ 8 ] due to the lack of haptic feedback and compromised eye-hand coordination [ 25 , 15 ] . Consequently, many methods incorporate haptic feedback by using physical keyboards [ 16 , 25 , 38 , 39 ] or other haptic devices [ 10 , 17 ] to enhance the differentiation between keys and improve selection accuracy. However, the dependency on additional hardware such as a physical keyboard restrict their usage to specific scenarios, such as when the user is seated at a desk where the keyboard can be placed, limiting their practicality in non-stationary VR experiences [ 41 ] . Therefore, most existing VR applications rely on virtual keyboards for text entry, and optimizing the usability of virtual keyboard-based entry has been an important topic [ 21 , 27 , 45 , 29 ] . In addition to usability, the security aspect of text entry in VR has also attracted increased attention in recent years. While passwords remain the most common form of sensitive input, other types of sensitive information, such as personal details, medical records, and confidential data from corporations or government agencies, are frequently entered via natural language [ 43 ] . Compromise of such natural language text entry can lead to serious consequences, ranging from scams and financial losses to threats to national security [ 43 ] , and a secure text entry method is a prerequisite for VR to be adopted in such contexts. Research has shown that the deterministic spatial relationships of the keys can be exploited, allowing effective prediction of the keyboard layouts and keystrokes. This issue is particularly pronounced in VR due to slower text entry rates, larger required body movements, reduced environmental awareness, and the presence of built-in sensor tracking, distinguishing it from traditional text entry scenarios such as those on smartphones or physical keyboards. For example, Ling et al. [ 28 ] proposed a keylogging inference attack that utilizes the orientation data of the headset and controller during text entry. These orientations can be inferred from recorded videos using computer vision algorithms or directly extracted from the controller’s sensor data. Similarly, Wu et al. [ 47 ] introduced a two-step keystroke snooping method. This method first reconstructs the virtual keyboard using onboard sensor data and then predicts the entered content. Their experiments demonstrated an 80–90% accuracy in recovering text or passwords. Similar keystroke inference attacks in VR have been widely discussed in other work, including [ 2 , 34 , 31 , 13 , 26 , 54 ] , etc. To counteract this vulnerability, research has focused on introducing randomization to disrupt these predictable spatial patterns, but naively randomizing keyboard layouts often leads to poor usability, as users struggle with unfamiliar and inconsistent designs [ 33 ] . Consequently, the primary challenge in addressing the security problem lies in implementing randomization strategies that minimize negative impacts on usability, but existing work that specifically considers the security challenge in VR text entry is limited [ 43 , 44 ] . In this paper, we propose a novel radial keyboard to address the two challenges discussed above, i.e., improving precise selection and enhancing security of text input through non-intrusive randomization. The specific contributions of this paper include: 1. We present a keystroke prediction strategy to illustrate the vulnerability of natural language text entry with limited randomization [ 43 , 44 ] and demonstrate its effectiveness against the state-of-the-art [ 43 ] . 2. We propose a secure virtual radial keyboard that is not only robust against keystroke snooping attack but also relaxes the precision requirement of key selection by resizing the keys in a random but non-intrusive fashion. 3. We conducted a user study to compare the proposed radial keyboard with the state-of-the-art [ 43 ] , and standard QWERTY keyboard in terms of both usability and security. 1 Related Work 1.1 Typing Methods for Virtual Keyboards Various typing methods have been explored for virtual keyboard input. According to [ 44 ] , mainstream typing techniques in VR can be categorized into position-based and ray-based approaches. With position-based text entry, users interact directly with the virtual keyboard by tapping or swiping on the keys using either their controllers [ 7 , 42 , 3 , 4 , 8 , 50 , 1 ] or their hands [ 17 , 46 , 9 , 1 ] . This method feels natural and intuitive, as it closely mimics how people interact with physical objects in the real world. However, position-based text entry tends to be relatively slow and can cause fatigue due to the extensive movements required, particularly when the keyboard is positioned mid-air [ 41 ] . In this paper, we focus on ray-based text entry using controllers. With ray-based text entry, users select keys by pointing at them with a virtual ray, typically controlled by a controller [ 7 , 44 , 43 ]