Large red dots indicate the locations of layouts F5 (in upper right diagrams) and F10 (in lower left diagram). FingerText: Mobile, One-Handed, Portable Text Input InProceedings of the 2021 CHI Conference on Human Factors in Computing Systems (CHI '21).
Wearable Technology
Inputs for Wearables
From these categories, this dissertation will only cover the direct manipulation of wearable devices; primary tasks such as context awareness are not addressed here. Accordingly, this dissertation will only refer to the smartwatch and smartglass design.
Contexts, Considerations and Design Factors
To avoid this unnecessary attention, the design of the wearable input should pursue discrete and socially acceptable actions, such as microinteractions [27]. In particular, a highly accessible system is needed for situational impairments, such as attention-demanding tasks [9], mobile use [4], and hand-occupied tasks [3, 33].
One-Handed, Intra-Hand Inputs
These different touch and motion actions have a wide input space of intra-hand input for diverse applications. In this thesis, the intra-hand input was implemented by the system with touch-sensitive fingernails and a motion sensor (Inertial Measurement Unit; IMU) on the wrist.
Claim, Structure, Contributions and Conclusion
This chapter takes advantage of this by using intra-hand input sensors for hand-to-face input. To understand intra-hand input and design such a system, this chapter reviews the general background in three sections.
Intra-Hand Inputs
Sensing Techniques
57] presented a system that can perform text input via a touch sensor pad on the tip of the index finger. In addition to these conventional approaches, many researchers have focused on bio-sensing techniques that use bio-acoustic or electronic signals generated while performing gestures.
Design Factors in Current Sensing Approaches
These applications take advantage of the properties of the eye, which is always available and very fast. On the other hand, although wearables are a new form factor for computers that are always on the body, there is a lack of consideration for the use of the human body itself in the current input modalities [99].
Touch-Sensitive Fingernails
However, such a decoupling may be associated with difficulties in creating an easy-to-understand mapping between user action and the input [98]. In this case, the near-body area of wearables can be used as an input surface, making it easier to create easy-to-understand maps.
Wearable Computing Contexts and its Considerations
For example, the use of wearable devices in public contexts, such as meetings, coffee shops or public transport, requires data that will not attract excessive attention from the public. Finally, Chapter 5 examines the use of such a system in a different way, namely hand-to-face, in terms of social acceptability.
Intra-hand Input in Commercial Wearables
Thus, future design will enable multiple hand input actions, such as pinch to each finger, flick, swipe, or poses [39] to provide sufficient input space for the needs of different applications. Notably, discrete inputs, such as tapping, can exploit wider input spaces when combined with continuous inputs, such as hand 3D movement.
Summary
Abstract
Design Factors in Nailz
Introduction
We contribute to research in this emerging field by proposing data via a set of five touch-sensitive nails - see Figure 1. We argue that data on how a set of touch-sensitive nails can be used to sense the articulation of hands is a valuable complement to previous work proposing nails as a cursor control surface [14, 115].
Related Work
Input on the Skin of the Hand
34] captures how comfortable this input type of input is, emphasizing that while touching the thumb to the inner surfaces of the index and middle fingers is relatively easy to perform, touch to the ring and little fingers is more taxing. 34] also report selection accuracy for different numbers of targets distributed over the length of the fingers; optimal input accuracy ranges from between two (on the little finger) and five (on index/middle) separate targets.
Input on Finger Nails
Finally, systems capable of detecting pose can capture entire hand gestures, such as those involved in sign language [67]. This paper extends this previous work by considering the full set of five nails instead of a single thumbnail, by investigating the inputs related to full hand articulation (instead of index finger control), and by providing a thorough description of performance covering comfort, timing, and accuracy.
Ideation Workshop
For example, moving the index finger from the thumb was proposed for sending a message, since the same action could remove a physical object. Similarly, moving the thumb over all the fingernails when they were vertically aligned was proposed for movement.
Initial Nail Input Set
Comfort Study
Participants and Method
The experiment lasted approximately 30 minutes and the participants were compensated the local equivalent of five USD. The study had participants try and then rate the ease of hand gestures shown to them on a laptop screen.
Results
This analysis indicates that for the nail variable, single nail touches of the thumb were significantly more pleasant than for other fingers and touches of the little finger were significantly less pleasant. Multifinger inputs involving the little finger were also rated as significantly less pleasant than those involving two or three of the other fingers.
Nail Sensor System
A candidate explanation for this is the relatively small scale of the movements involved in nail touches, compared to the stretching required to touch areas such as the proximal phalanxes of the fingers. All five spikes in a set were wired to a single wrist-mounted Arduino MKR1010 with lightweight AWG32 wires that did not restrict finger movements.
Performance Study
- Participants, Procedure and Design
- Nail Input Recognizer
- Distinguishing Input Actions
- Time Results
We created a simple decision tree to classify fingernail taps from five sensors into one of 144 fingernail inputs. We performed a final RM-ANOVA on the temporal data from each of the seven actions involving the joint center region.
Revised System Design
Final Input Set
Revised Recognizer
Sample Applications
Users can play/pause content by tapping their index finger and middle nail, a configuration where the pair of fingers resembles a pause icon. For continuous input, you can combine fingernail touches with motion data: tap and hold the tip of the index finger while rotating your hand to control playback position.
Verification Study
Participants and Method
In the final phase, we showed the two applications described in the previous section to the participants, let them try and experience them for 10 minutes, and then conducted a semi-structured interview to capture their reactions and opinions. In total, this study retained 870 first-phase surveys, 870 second-phase surveys, and approximately 60 minutes of transcribed interview content.
Results
The comfort and usefulness of some of the input actions was questioned, especially by three participants who emphasized “input on small nail is frustrating” (P0, P5, P7), possibly due to Midas touches because “the little finger is bent, so little nail was touched unintentionally” (P9). It was felt that some of the input actions were intended to alleviate this problem: P0 suggested that long tapping to open "can prevent errors".
Discussion and Conclusion
These show that nail taps can be performed quickly - at an average of 1.61 s in the final phase of the validation study, a figure lower than the 1.32 to 2.448 s reported for taps caused by landing actions and removal by Lee et al. We show that fingertip inputs are expressive, can be performed comfortably and efficiently, and are easily recognized using simple criteria. We believe that the range of small-scale inputs supported, the simple classification scheme, and the speed with which participants perform all demonstrate the robustness of this approach.
Abstract
Design Factors in FingerText
These situations require portable text input systems that can be operated with one hand while engaging in common activities such as walking. The simplicity and easy availability of this modality has also led to diverse text input designs.
Related Work
Wearable Text Entry
Text input systems based on the touches between the fingers of one hand are particularly relevant to this article. This finger region has not yet been studied in the context of text input, and we have identified a number of reasons why it may be particularly suited to this form of input.
Mobile Input
First, nails have been associated with improved comfort ratings [154] compared to those reported on finger phalanges. On the other hand, we argue that the intra-hand touches we are studying and based on relatively large movements of all fingers relative to each other can be very resistant to walking impacts [153].
Keyboard Optimization
The studies in this article are the first to investigate portable one-handed text input while on mobile. Walking is likely to disrupt motion input, increase the impact of annoyance [166] and disrupt the acquisition of small targets [167], factors that indicate that the ability of these previously proposed techniques to support reliable and effective input while users are in transit. on the road can be very limited.
Experimental Platform
Regarding feedback, all studies reported in this article used an Epson BT 200 Head Mounted Display (HMD) to present instructions and interfaces to users. All visual content on this device was managed by the host computer over a wireless UDP connection.
Metrics Study
- Design
- Participants
- Measures
- Procedure
- Results and Discussion
- Conclusion
These results suggest that the most important factor affecting performance was the relationship between the initial and final touch finger areas; the interactions led to the largest effect sizes in the study. Rather than depending on which finger is touched, speed (and possibly accuracy) in the dual-touch task we studied depended on the relationship between the start and end points.
Keyboard Layout Optimization
- Metrics
- Speed
- Accuracy
- Comfort
- Confusability
- Normalization
- Multi-objective Optimization
- Results and Layout Selection
We limited the minimum number of letters that could be assigned to each key in both F5 and F10 to one and the maximum to six and four respectively. We used the same platform (NSGA-II/Pymoo) to perform multiple objective optimizations using the normalized metrics for both F5 and F10 formats.
Text Entry Study
- Design
- Participants
- Measures
- Procedure
- Results
- Discussion
In particular, in the case of the Qwerty keyboards, additional explanations were given to ensure that the participants were familiar with the layout. In addition, we saw little benefit of QWERTY layouts in early trials in either block—initial performance and learning curves were not significantly better.
Text Entry System Design
Instead of writing on the small screen of your watch, it may be better to simply type by hand [44, 57].
Limitations
Conclusion
Appendix: Optimization Process
- Objective Function
- Speed
- Accuracy
- Comfort
- Confusability
Terms are identical to equation (2) saw forakl, defined as the average accuracy when touching the nail region after touching the nail region k.
Abstract
Design Factors in Hand-to-Face Input
These objective capabilities indicate that inputs can be made directly and quickly.
Introduction
We contribute both this new combination of elicitation and social acceptance methods and the research findings in terms of strategies for designing socially acceptable hand-to-face input techniques for AR/wearables. In this way, we demonstrate the value of our design strategies to create socially acceptable hand-to-face input techniques for AR/wearables.
Related Work
Subtlety, Unobtrusiveness and Social Acceptability
Based on these ideas, social acceptability has been defined as whether an input action appears appropriate to both the user performing it and the observers observing it in the context in which it occurs [221]. 221] provide a useful classification of the social acceptability of gestures based on whether the input actions and/or the resulting outcomes are observable.
Hand-to-Face Elicitation Study
Experimental Design, Tasks and Setting
Participants were also instructed to generate discrete or subtle actions suitable for use in the public setting of the study, and both they and their partners rated all preferred input propositions for social acceptability and how obvious the input action was. For obviousness (or discreetness), we adapted the social acceptability questionnaire to ask in which places an action would be obvious.
Participants and Procedure
In total, participants answered four questions on the favorite input action that both they and their partner had produced. After both participants created two suggestions in pairs and chose their favorite, they rated their social acceptability and how obvious they were.
Results and Data Processing
- Generation Strategies
Similarly, actions such as flexion, which is universally used for a deviated ear, scored highly. Concealment is the use of an area of the face that is naturally hidden from observers in front of the user, such as the back of the neck or, when the head is turned, the ear.
Case Studies
Case Study 1 - EarTouch
- Performance Study
- Results and Discussion
In such situations, the view of the ear is static, and the background is moving rapidly. With these techniques, we have achieved a simple, reliable and stable model of the shape of the ear.
Case Study 2 - ThumbTouch
- Performance Study
- Results and Discussion
This is likely due to the contact of the entire edge of the nail with the face. At the end of the study, we conducted a semi-structured interview with the participant couple capturing their opinions and reactions to the given techniques.
Experiment Results
The value of miniaturization was also seen in an assessment of input techniques – larger or prolonged movements in the pan tasks were “strange” and would “attract attention” (U1) and actions involving larger physical movements such as "taking the finger". outside the ear" in liftOff (O2) stood out. Participants also suggested this strategy for new designs: under the chin, on the neck or behind the ear would "hide the gesture well" (O6).
Conclusions
In particular, the interview data reveal that both users and observers, who respectively performed and naively watched input tasks on the prototypes, often highlighted aspects of the strategies as important qualities or mechanisms for achieving socially acceptable and non-obvious or discrete hand-to-face input. The next steps for EarTouch involve integrating machine learning algorithms to detect a larger number of hand actions, such as bends, pushes or deformations of the ear.
Summary
Design Factors in the Three Scenarios
- Manipulation
- Accessibility
- Space
- Social
Chapter 3: The Ideation Workshop generated various input primitives within the hand (tapping, flicking, and sliding across multiple nail regions) and derived a large set of 144 input actions from these input primitives. Compared to these works, this study explored more different in-hand insertion actions within a nail by examining multiple actions (tapping, sliding, and flicking), nail combinations (single and multiple nails), nail regions (tip, center, root and both.
Considerations
- Tracking Technologies
- Body Regions
- Input Types
- Design Processes
Regarding the acceptability score presented in Chapter 5, the highest score was on the ear, neck, and temple, suggesting that input areas away from the center of the face may be appropriate for hand-to-face design. First, to improve the ecological validity of the proposed actions, the study was conducted in a busy public place (a coffee shop).
Lesson Learned, Limitation and Future Work
- Multimodal Input
- Interface Design
- Customization and Bespoke Design
- In the Wild: Midas Touch
- Aesthetic
In particular, the custom design of the input device can affect comfort and input performance [7, 38]. Firstly, the custom design of the input device can prevent such accidental touches as mentioned above.
Practical Guidance
- Manipulation
- Accessibility
- Space
- Social
In addition, multimodal input can also increase the input space to a large extent. In fact, a proper planning process should be adopted to better explore the entrance space.
Conclusion
In Proceedings of the 12th International Conference on Human Computer Interaction with Mobile Devices and Services, pp. In Proceedings of the 10th International Conference on Human Computer Interaction with Mobile Devices and Services, pp.
