Slide presentations are central to classrooms and conferences but remain inaccessible to blind and low vision (BLV) audiences. Presenters rarely describe visual content or announce slide changes, leaving BLV participants with fragmented access. We present a delay-buffered editing approach that trims redundant speech and inserts concise slide descriptions at transition points in recorded presentations, operating within a five‑second buffer. An exploratory study showed that trimming filler speech created sufficient space for descriptions and that added descriptions improved comprehension and orientation. Building on these findings, we developed an automated editing pipeline that processes recorded presentation videos and evaluated it with 12 BLV participants. Edited audios improved comprehension (from 50.0% to 88.9%), slide detection (from 16.7% to 100%), and recognition of visual elements, while participants also noted challenges in timing and prosody. These results establish delay-buffered editing as a promising approach for enhancing accessibility of recorded presentations with slides, and suggest design directions for future live deployment.

Catala is a domain-specific programming language for statutory law, featuring prioritized default logic and programming constructs that precisely mirror legal reasoning structures. Extensive study of block-based programming has not clarified how such representations function in semantics-constrained domain-specific languages that require users to directly encode statutory logic. We introduce CatalaBlocks, a block-based representation of Catala that renders its core semantic constructs into a constrained visual form, enabling empirical study of how representation shapes interaction with computationally formalized statutory logic. We conducted a comparative study with legal professionals, examining how participants implemented statutory rules using either textual Catala or CatalaBlocks. Participants using CatalaBlocks completed tasks more quickly, produced more accurate code, reported lower difficulty, and expressed greater confidence in the solutions’ alignment with intended statutory logic. These findings provide insight into how representation shapes domain experts’ interaction with a semantics-first language for statutory law.

Live presentations often pose accessibility challenges for blind and low vision (BLV) audiences due to limited verbal references to slide content.We propose a novel approach that introduces a short temporal buffer to the live stream, enabling the removal of redundant speech-such as filler phrases, repetitions, and side remarks-and the insertion of concise slide descriptions in the resulting gaps.To explore the effectiveness of this method, We conducted a Wizard-of-Oz-inspired study using pre-recorded videos that simulate real-time editing using edited presentation videos that included slide numbers, titles, and image descriptions.Four blind participants viewed both edited and unedited versions and shared their experiences in post-task interviews.Participants reported clearer transitions, better comprehension of slide content, and improved ability to mentally visualize slide layouts in the edited condition.We also analyze the types of removable speech and outline key design and technical challenges in building a real-time, automated version of the system.Our findings highlight the potential of near-real-time editing as a lightweight strategy to enhance the accessibility of live spoken content.While the study simulates live accessibility using pre-recorded content, it provides valuable insight into future real-time applications.

Object recognition technologies hold the potential to support blind and low-vision people in navigating the world around them. However, the gap between benchmark performances and practical usability remains a significant challenge. This paper presents a study aimed at understanding blind users' interaction with object recognition systems for identifying and avoiding errors. Leveraging a pre-existing object recognition system, URCam, fine-tuned for our experiment, we conducted a user study involving 12 blind and low-vision participants. Through in-depth interviews and hands-on error identification tasks, we gained insights into users' experiences, challenges, and strategies for identifying errors in camera-based assistive technologies and object recognition systems. During interviews, many participants preferred independent error review, while expressing apprehension toward misrecognitions. In the error identification task, participants varied viewpoints, backgrounds, and object sizes in their images to avoid and overcome errors. Even after repeating the task, participants identified only half of the errors, and the proportion of errors identified did not significantly differ from their first attempts. Based on these insights, we offer implications for designing accessible interfaces tailored to the needs of blind and low-vision users in identifying object recognition errors.

Teachable object recognizers provide a solution for a very practical need for blind people – instance level object recognition. They assume one can visually inspect the photos they provide for training, a critical and inaccessible step for those who are blind. In this work, we engineer data descriptors that address this challenge. They indicate in real time whether the object in the photo is cropped or too small, a hand is included, the photos is blurred, and how much photos vary from each other. Our descriptors are built into open source testbed iOS app, called MYCam. In a remote user study in (N = 12) blind participants’ homes, we show how descriptors, even when error-prone, support experimentation and have a positive impact in the quality of training set that can translate to model performance though this gain is not uniform. Participants found the app simple to use indicating that they could effectively train it and that the descriptors were useful. However, many found the training being tedious, opening discussions around the need for balance between information, time, and cognitive load.

Negative attitudes shape experiences with stigmatized conditions such as dementia, from affecting social relationships to influencing willingness to adopt technology. Consequently, attitudinal change has been identified as one lever to improve life for people with stigmatized conditions. Though recognized as a scaleable approach, social media has not been studied in terms of how it should best be designed or deployed to target attitudes and understanding of dementia. Through a mixed methods design with 123 undergraduate college students, we study the effect of being exposed to dementia-related media, including content produced by people with dementia. We selected undergraduate college students as the target of our intervention, as they represent the next generation that will work and interact with individuals with dementia. Our analysis describes changes over the period of two weeks in attitudes and understanding of the condition. The shifts in understanding of dementia that we found in our qualitative analysis were not captured by the instrument we selected to assess understanding of dementia. While small improvements in positive and overall attitudes were seen across all interventions and the control, we observe a different pattern with negative attitudes, where transcriptions of content produced by people with dementia significantly reduced negative attitudes. The discussion presents implications for supporting people with dementia as content producers, doing so in ways that best affect attitudes and understanding by drawing on research on cues and interactive media, and supporting students in changing their perspectives towards people with dementia.

Teachable interfaces can empower end-users to attune machine learning systems to their idiosyncratic characteristics and environment by explicitly providing pertinent training examples. While facilitating control, their effectiveness can be hindered by the lack of expertise or misconceptions. We investigate how users may conceptualize, experience, and reflect on their engagement in machine teaching by deploying a mobile teachable testbed in Amazon Mechanical Turk. Using a performance-based payment scheme, Mechanical Turkers (N=100) are called to train, test, and re-train a robust recognition model in real-time with a few snapshots taken in their environment. We find that participants incorporate diversity in their examples drawing from parallels to how humans recognize objects independent of size, viewpoint, location, and illumination. Many of their misconceptions relate to consistency and model capabilities for reasoning. With limited variation and edge cases in testing, the majority of them do not change strategies on a second training attempt.

For people with visual impairments, photography is essential in identifying objects through remote sighted help and image recognition apps. This is especially the case for teachable object recognizers, where recognition models are trained on user's photos. Here, we propose real-time feedback for communicating the location of an object of interest in the camera frame. Our audio-haptic feedback is powered by a deep learning model that estimates the object center location based on its proximity to the user's hand. To evaluate our approach, we conducted a user study in the lab, where participants with visual impairments (N=9) used our feedback to train and test their object recognizer in vanilla and cluttered environments. We found that very few photos did not include the object (2% in the vanilla and 8% in the cluttered) and the recognition performance was promising even for participants with no prior camera experience. Participants tended to trust the feedback even though they know it can be wrong. Our cluster analysis indicates that better feedback is associated with photos that include the entire object. Our results provide insights into factors that can degrade feedback and recognition performance in teachable interfaces.

Speech has become an increasingly common means of text input, from smartphones and smartwatches to voice-based intelligent personal assistants. However, reviewing the recognized text to identify and correct errors is a challenge when no visual feedback is available. In this paper, we first quantify and describe the speech recognition errors that users are prone to miss, and investigate how to better support this error identification task by manipulating pauses between words, speech rate, and speech repetition. To achieve these goals, we conducted a series of four studies. Study 1, an in-lab study, showed that participants missed identifying over 50% of speech recognition errors when listening to audio output of the recognized text. Building on this result, Studies 2 to 4 were conducted using an online crowdsourcing platform and showed that adding a pause between words improves error identification compared to no pause, the ability to identify errors degrades with higher speech rates (300 WPM), and repeating the speech output does not improve error identification. We derive implications for the design of audio-only speech dictation.

Precisely guiding a blind person's hand can be useful for a range of applications from tracing printed text to learning and understanding shapes and gestures. In this paper, we evaluate wrist-worn haptics as a directional hand guide. We implemented and evaluated the following haptic wristband variations: (1) four versus eight vibromotor designs; (2) vibration from only a single motor at a time versus from two adjacent motors using interpolation. To evaluate our designs, we conducted two studies: Study 1 (N=13, 2 blind) showed that participants could non-visually find targets and trace paths more quickly and accurately with single-motor feedback than with interpolated feedback, particularly when only four motors were used. Study 2 (N=14 blind or visually impaired participants) found that single-motor feedback with four motors was faster, more accurate, and most preferred compared to similar feedback with eight motors. We derive implications for the design of wrist-worn directional haptic feedback and discuss future work.

Current symbol-based dictionaries providing vocabulary support for persons with the language disorder, aphasia, are housed on smartphones or other portable devices. To employ the support on these external devices requires the user to divert their attention away from their conversation partner, to the neglect of conversation dynamics like eye contact or verbal inflection. A prior study investigated head-worn displays (HWDs) as an alternative form factor for supporting glanceable, unobtrusive, and always-available conversation support, but it did not directly compare the HWD to a control condition. To address this limitation, we compared vocabulary support on a HWD to equivalent support on a smartphone in terms of overall experience, perceived focus, and conversational success. Lastly, we elicited critical discussion of how each device might be better designed for conversation support. Our work contributes (1) evidence that a HWD can support more efficient communication, (2) preliminary results that a HWD can provide a better overall experience using assistive vocabulary, and (3) a characterization of the design features persons with aphasia value in portable conversation support technologies. Our findings should motivate further work on head-worn conversation support for persons with aphasia.

The QWERTY keyboard has been a de facto standard for computer text entry and continues to be one for mobile text entry such as for smartphones. It is not clear, however, that it will continue to be an option for text entry for much smaller devices such as smartwatches. In a series of user experiments, we examined the performance of the QWERTY keyboard when it is reduced to fit a small smartwatch screen. At the same time, we examined whether the ZoomBoard and the SplitBoard, which are QWERTY keyboards augmented by zooming and panning strategies, respectively, would be effective in comparison with a plain QWERTY keyboard. In Experiment 1, we evaluated the text entry performance of new users on the three QWERTY keyboards. In Experiment 2, we evaluated the relative performance of the three keyboards for three different screen sizes. In Experiment 3, we further observed how the keyboard performance changed when used in a mobile situation. Main results are: (i) users could adapt to a plain QWERTY keyboard even in the smallest screen cases. (ii) The SplitBoard consistently showed a better performance than other keyboards in all tested sizes. (iii) The SplitBoard showed a better performance than other keyboards in a mobile condition (treadmill) and was preferred most by participants.

The two main tasks of a smart TV GUI are menu navigation and free pointing. Traditional remotes with directional keys are suitable for menu navigation but may not be so for free pointing. More recent remotes with a two-dimensional (2D) pointing device are suitable for free pointing but may not be so for menu navigation. To support both types of tasks well, we devised a new input device called TouchRoller. We expect that it can support both types of tasks well because it has a separable control structure and a continuous input property. A comparative user study showed that the performance of TouchRoller is comparable to that of directional keys for menu navigation and 2D pointing devices for free pointing. In addition, it was most favored by the participants, and NASA TLX test results showed that TouchRoller demands the lowest task load.

Text entry on a smartwatch is a challenging problem due to the device's limited screen area. In this paper, we introduce the SplitBoard, which is a soft keyboard designed for a smartwatch. As the user flicks left or right on the keyboard, it switches between the left and right halves of a QWERTY keyboard. We report the results of two user experiments where the SplitBoard was compared to an ordinary QWERTY keyboard, the ZoomBoard, SlideBoard, and Qwerty-like keypad. We measured the initial performance with new users for each method. The SplitBoard outperformed all other techniques in the experiments. The SplitBoard is expected to be a viable option for smartwatch text entry because of its light processing requirements, good performance, and immediate learnability.