TAGSwipe: Touch Assisted Gaze Swipe for Text Entry

Kumar, Chandan; Hedeshy, Ramin; MacKenzie, I. Scott; Staab, Steffen

doi:10.1145/3313831.3376317

Cited by 31 publications

(18 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…EyeSwipe can achieve a typing speed of 11.7 wpm after 30 minutes of typing. Kumar et al [37] presented ''TAGSwipe,'' a gaze and touchbased system mapping gaze path into words and can achieve a typing speed of 15.45 wpm.…”

Section: B Accessible Interactionsmentioning

confidence: 99%

“…From all the related work discussed on gaze-assisted interactions, it can be observed that they use dwell-based activation, specific gaze gestures, or smooth pursuit movement of the gaze to trigger dedicated actions. Applications like ''TAGSwipe [37],'' gaze gesture-based authentication [10], [11], gaze gesture guiding system [31], and so on did primarily use gaze gestures to express a user's intended action; however, the gestures used and the recognition method remain application-specific. This lack of a generic gesture recognition framework motivated us to explore gaze gesture design strategies, recognition algorithms, and time and performance measures of various recognition algorithms.…”

Section: E Entertainmentmentioning

confidence: 99%

See 1 more Smart Citation

Evaluating the Performance of Machine Learning Algorithms in Gaze Gesture Recognition Systems

Ray

Rajanna

et al. 2022

IEEE Access

View full text Add to dashboard Cite

Despite the utility of gaze gestures as an input method, there is a lack of guidelines available regarding how to design gaze gestures, what algorithms to use for gaze gesture recognition, and how these algorithms compare in terms of performance. To facilitate the development of applications that leverage gaze gestures, we have evaluated the performance of a combination of template-based and data-driven algorithms on two custom gesture sets that can map to user actions. Template-based algorithms had consistently high accuracies but the slowest runtimes, making them best for small gesture sets or accuracy-critical applications. Data-driven algorithms run much faster and scale better to larger gesture sets, but require more training data to achieve the accuracy of the template-based methods. The main takeaways for gesture set design are 1) gestures should have distinct forms even when performed imprecisely and 2) gestures should have clear key-points for the eyes to fixate onto.INDEX TERMS Eye tracking, gaze gestures, gesture design, human-computer interaction, multimodal interaction.

show abstract

Section: B Accessible Interactionsmentioning

confidence: 99%

Section: E Entertainmentmentioning

confidence: 99%

Evaluating the Performance of Machine Learning Algorithms in Gaze Gesture Recognition Systems

Ray

Rajanna

et al. 2022

IEEE Access

View full text Add to dashboard Cite

show abstract

“…The naive solution to the Midas problem [2] is using dwell to indicate the intent of selection confirmation. Nevertheless, "gaze pointing + dwell confirmation" is slow and uncomfortable [5]. Sidenmark et al [19] proposed to utilize the distinction between gaze shift performed only by eyes and head-supported gaze shift, i.e., eye movement accompanied by head movement, to enable hover interaction, visual exploration around pre-selected targets, etc.…”

Section: Bi-modal Target Selection Involving Eye Gazementioning

confidence: 99%

“…To address the Midas problem, despite the naive solution dwell that is reported to be slow and uncomfortable [5], a secondary interaction modality for target selection confirmation (e.g., using two fingers to make an air-tap gesture, pressing a button on a device, etc.) is usually combined with eye gaze pointing.…”

Section: Introductionmentioning

confidence: 99%

Investigating the Effects of Mapping Gaze-selected Sub-menus to a Cellphone Touchscreen on Target Selection in AR

Tian¹,

Zheng²,

Zhao³

2022

Preprint

View full text Add to dashboard Cite

<p>Eye gaze is potentially fast and ergonomic for target selection in AR. Nevertheless, it is reported to be inaccurate. To compensate for its low accuracy in selecting targets in an AR menu, previous researchers proposed dividing a menu into several sub-menus where targets are arranged regularly, and mapping the one pointed by eye gaze to a Google Glass’ touchpad on which the user confirms the selection of a target in the sub-menu via swipe or tap. However, whether this technique was effective in enhancing the basic gaze-touch target selection was not investigated. We coined the terminology TangibleCounterpart for this technique's essence and suggested using a cellphone touchscreen as the “tangible counterpart” for sub-menus. Further, we proposed the design space of the cellphone-based TangibleCounterpart concept that includes three dimensions, i.e., sub-menu size, the way a user holds the cellphone, and the touch technique for selection confirmation. Our empirical study showed that only when participants used two thumbs for selection confirmation (TH) and the sub-menu size was 2 rows * 1 column (2R1C) or 2 rows * 2 columns (2R2C), the error rates were significantly smaller than that under the basic gaze-touch target selection; and speeds were not compromised. The best performance (error rate: mean = 4%, SD = 2.6%; completion time: mean = 1.2s, SD = 0.2s) was achieved when participants used tapping for selection confirmation under TH and 2R2C. </p>

show abstract

“…In recent years, several commercial AR and VR HMDs have introduced integrated eye tracking functionality. Ahn et al [3] and Kumar et al [56] combined gaze and touch to input text and Rajanna et al [78] combined gaze and a button click. The additional touch modality is used to select a key, and speeding up eye-tracking for text entry, usually using dwell time over a key to confirm inputs.…”

Section: Gaze Based Text Entrymentioning

confidence: 99%

Extended Reality for Knowledge Work in Everyday Environments

Biener¹,

Ofek²,

Pahud³

et al. 2021

Preprint

View full text Add to dashboard Cite

Virtual and Augmented Reality (VR and AR) have the potential to change information work. The ability to modify the workers senses can transform everyday environments into a productive office, using portable head-mounted displays (HMDs) combined with conventional interaction devices, such as keyboards and tablets. While a stream of better, cheaper and lighter HMDs have been introduced for consumers in recent years, there are still many challenges to be addressed to allow this vision to become reality. This chapter summarizes the state of the art in the field of extended reality for knowledge work in everyday environments and proposes steps to address the open challenges.

show abstract

TAGSwipe: Touch Assisted Gaze Swipe for Text Entry

Cited by 31 publications

References 44 publications

Evaluating the Performance of Machine Learning Algorithms in Gaze Gesture Recognition Systems

Evaluating the Performance of Machine Learning Algorithms in Gaze Gesture Recognition Systems

Investigating the Effects of Mapping Gaze-selected Sub-menus to a Cellphone Touchscreen on Target Selection in AR

Extended Reality for Knowledge Work in Everyday Environments

Contact Info

Product

Resources

About