SliceType: Fast Gaze Typing with a Merging Keyboard (1706.02499v4)

Abstract: Jitter is an inevitable by-product of gaze detection. Because of this, gaze typing tends to be a slow and frustrating process. In this paper, we propose SliceType, a soft keyboard that is optimized for gaze input. Our main design objective is to use the screen area more efficiently by allocating a larger area to the target keys. We achieve this by determining the keys that will not be used for the next input, and allocating their space to the adjacent keys with a merging animation. Larger keys are faster to navigate towards, and easy to dwell on in the presence of eye tracking jitter. As a result, the user types faster and more comfortably. In addition, we employ a word completion scheme that complements gaze typing mechanics. A character and a related prediction is displayed at each key. Dwelling at a key enters the character, and double-dwelling enters the prediction. While dwelling on a key to enter a character, the user reads the related prediction effortlessly. The improvements provided by these features are quantified using the Fitts' law. The performance of the proposed keyboard is compared with two other soft keyboards designed for gaze typing, Dasher and GazeTalk. 37 novice users gaze-typed a piece of text using all three keyboards. The results of the experiment show that the proposed keyboard allows faster typing, and is more preferred by the users.

• The paper presents a dynamic keyboard that merges less likely keys based on predictive language models, significantly enhancing gaze typing speed and accuracy.
• It integrates on-key word predictions to minimize gaze shifts, effectively reducing cognitive load during text entry.
• Experimental validation with 37 participants showed that SliceType outperforms alternatives like Dasher and GazeTalk, confirming its intuitive design and efficiency.

Review of "SliceType: Fast Gaze Typing with a Merging Keyboard"

The paper "SliceType: Fast Gaze Typing with a Merging Keyboard" by Burak Benligiray, Cihan Topal, and Cuneyt Akinlar introduces an innovative soft keyboard specifically optimized for gaze input. This research is grounded on addressing the inherent issues associated with gaze typing, primarily eye-tracking jitter that slows down text entry and reduces accuracy. The authors propose the SliceType keyboard, which intelligently allocates screen space for the keys most likely to be used next, effectively enlarging potential target areas and thereby enhancing gaze typing efficiency.

Methodology and Key Innovations

The authors' primary methodological contribution is the design of SliceType, which increases typing speed by dynamically adjusting key sizes based on predicted next inputs. This approach is grounded in the usage of a predictive LLM that assists in identifying and excluding improbable next keys, subsequently merging their space with neighboring keys. Such dynamically adjustable key sizing potentially reduces the time needed to navigate between keys and counters the challenges posed by gaze tracker inaccuracy.

Key features of the SliceType design include:

1. Dynamic Key Merging:
   • Keys that are less probable to be used next, based on language predictions, are temporarily removed, and their space is merged with adjacent keys. This design significantly increases the efficiency of gaze typing by expanding the size of the probable target keys.
2. Integrated Predictions:
   • Predictions are embedded within each key, allowing users to see word completions during their primary gaze dwell. This integrated approach minimizes the requirement for users to shift their focus to a separate area dedicated to word predictions, thus reducing additional cognitive load and gaze transitions.
3. Optimized Keyboard Layout:
   • The arrangement includes a central area with frequently used keys and an outer ring with less commonly used ones. This layered approach minimizes travel time and supports the merging function, as frequently used keys typically have an adjacent, less commonly used key.

Experimental Validation

SliceType's performance was evaluated against two other gaze-optimized keyboards, Dasher and GazeTalk, through a user paper involving 37 novice participants. The results illustrated that SliceType outperformed its counterparts in terms of typing speed. Participants showed a higher preference for SliceType, indicating its intuitive design and enhanced user-friendliness.

SliceType's effectiveness is quantitatively analyzed through Fitts' law, substantiating that key merging offers a more significant performance improvement compared to word completion. It demonstrates lower Index of Difficulty (ID) scores compared to traditional methods, implying a faster and more comfortable user experience in gaze typing contexts.

Implications and Future Directions

The research addresses critical aspects of gaze-based text entry systems, contributing to the field of assistive technology for individuals with motor impairments and applications where hands-free interaction is required. The notion of dynamically allocating interface space is valuable for reducing eye strain and increasing efficiency in user-interface design, especially in ubiquitous computing environments.

Future research directions could include enhancing prediction models using more complex language processing techniques or deep learning frameworks, which might offer improved prediction accuracy and adaptability to user behavior. Additionally, further exploration into the ergonomic and visual comfort factors for diverse user demographics might extend SliceType's applicability to broader contexts.

In summary, the paper presents a substantive contribution to the field by addressing the usability challenges in gaze-based typing through innovative interface design. The insights into dynamic key sizing and intuitive interaction open avenues for further improvements in gaze interaction technologies and beyond.