Leah Findlater

Colin Lea, Dianna Yee, Jaya Narain et al.

Many consumer speech recognition systems are not tuned for people with speech disabilities, resulting in poor recognition and user experience, especially for severe speech differences. Recent studies have emphasized interest in personalized speech models from people with atypical speech patterns. We propose a query-by-example-based personalized phrase recognition system that is trained using small amounts of speech, is language agnostic, does not assume a traditional pronunciation lexicon, and generalizes well across speech difference severities. On an internal dataset collected from 32 people with dysarthria, this approach works regardless of severity and shows a 60% improvement in recall relative to a commercial speech recognition system. On the public EasyCall dataset of dysarthric speech, our approach improves accuracy by 30.5%. Performance degrades as the number of phrases increases, but consistently outperforms ASR systems when trained with 50 unique phrases.

Yu-Yun Tseng, Tanusree Sharma, Lotus Zhang et al.

Individuals who are blind or have low vision (BLV) are at a heightened risk of sharing private information if they share photographs they have taken. To facilitate developing technologies that can help them preserve privacy, we introduce BIV-Priv-Seg, the first localization dataset originating from people with visual impairments that shows private content. It contains 1,028 images with segmentation annotations for 16 private object categories. We first characterize BIV-Priv-Seg and then evaluate modern models' performance for locating private content in the dataset. We find modern models struggle most with locating private objects that are not salient, small, and lack text as well as recognizing when private content is absent from an image. We facilitate future extensions by sharing our new dataset with the evaluation server at https://vizwiz.org/tasks-and-datasets/object-localization.

Neelima Prasad, Jarek Reynolds, Neel Karsanbhai et al.

We propose a novel task, hierarchical instance tracking, which entails tracking all instances of predefined categories of objects and parts, while maintaining their hierarchical relationships. We introduce the first benchmark dataset supporting this task, consisting of 2,765 unique entities that are tracked in 552 videos and belong to 40 categories (across objects and parts). Evaluation of seven variants of four models tailored to our novel task reveals the new dataset is challenging. Our dataset is available at https://vizwiz.org/tasks-and-datasets/hierarchical-instance-tracking/

Lucy Jiang, Amy Seunghyun Lee, Jon E. Froehlich et al.

Public art can hold cultural, social, political, and aesthetic significance, enriching urban environments and promoting well-being. However, a majority of urban art is inaccessible to blind and low vision (BLV) people. Most art access research has focused on private and curated settings (e.g., museums, galleries) and most urban access work has centered on outdoor navigation, leaving urban and public art accessibility largely understudied. We conducted semi-structured interviews with 16 BLV participants, using design probes featuring AI-generated descriptions and real-time AI interactions to investigate preferences for both discovering and engaging with urban art. We found that BLV people valued spontaneous art exploration, multisensory (e.g., tactile, auditory, olfactory) engagement, and detailed descriptions of culturally significant artwork. Participants also highlighted challenges distinct to urban art contexts: safety took precedence over art exploration, multisensory access measures could be disruptive to others in the public space, and inaccurate AI descriptions could lead to cultural erasure. Our contributions include empirical insights on BLV preferences for urban art discovery and engagement, seven design dimensions for public art access solutions, and implications for expanding HCI urban accessibility research beyond navigation.

Colin Lea, Vasileios Baltatzis, Connor Gillis et al.

AI-driven sign language interpretation is limited by a lack of high-quality annotated data. New datasets including ASL STEM Wiki and FLEURS-ASL contain professional interpreters and 100s of hours of data but remain only partially annotated and thus underutilized, in part due to the prohibitive costs of annotating at this scale. In this work, we develop a pseudo-annotation pipeline that takes signed video and English as input and outputs a ranked set of likely annotations, including time intervals, for glosses, fingerspelled words, and sign classifiers. Our pipeline uses sparse predictions from our fingerspelling recognizer and isolated sign recognizer (ISR), along with a K-Shot LLM approach, to estimate these annotations. In service of this pipeline, we establish simple yet effective baseline fingerspelling and ISR models, achieving state-of-the-art on FSBoard (6.7% CER) and on ASL Citizen datasets (74% top-1 accuracy). To validate and provide a gold-standard benchmark, a professional interpreter annotated nearly 500 videos from ASL STEM Wiki with sequence-level gloss labels containing glosses, classifiers, and fingerspelling signs. These human annotations and over 300 hours of pseudo-annotations are being released in supplemental material.

Tanusree Sharma, Yu-Yun Tseng, Lotus Zhang et al.

Blind and low vision (BLV) individuals use Generative AI (GenAI) tools to interpret and manage visual content in their daily lives. While such tools can enhance the accessibility of visual content and so enable greater user independence, they also introduce complex challenges around visual privacy. In this paper, we investigate the current practices and future design preferences of blind and low vision individuals through an interview study with 21 participants. Our findings reveal a range of current practices with GenAI that balance privacy, efficiency, and emotional agency, with users accounting for privacy risks across six key scenarios, such as self-presentation, indoor/outdoor spatial privacy, social sharing, and handling professional content. Our findings reveal design preferences, including on-device processing, zero-retention guarantees, sensitive content redaction, privacy-aware appearance indicators, and multimodal tactile mirrored interaction methods. We conclude with actionable design recommendations to support user-centered visual privacy through GenAI, expanding the notion of privacy and responsible handling of others data.

Griffin Dietz Smith, Dianna Yee, Jennifer King Chen et al.

Identifying mistakes (i.e., miscues) made while reading aloud is commonly approached post-hoc by comparing automatic speech recognition (ASR) transcriptions to the target reading text. However, post-hoc methods perform poorly when ASR inaccurately transcribes verbatim speech. To improve on current methods for reading error annotation, we propose a novel end-to-end architecture that incorporates the target reading text via prompting and is trained for both improved verbatim transcription and direct miscue detection. Our contributions include: first, demonstrating that incorporating reading text through prompting benefits verbatim transcription performance over fine-tuning, and second, showing that it is feasible to augment speech recognition tasks for end-to-end miscue detection. We conducted two case studies -- children's read-aloud and adult atypical speech -- and found that our proposed strategies improve verbatim transcription and miscue detection compared to current state-of-the-art.

Dhruv Jain, Khoa Huynh Anh Nguyen, Steven Goodman et al.

Recent advances have enabled automatic sound recognition systems for deaf and hard of hearing (DHH) users on mobile devices. However, these tools use pre-trained, generic sound recognition models, which do not meet the diverse needs of DHH users. We introduce ProtoSound, an interactive system for customizing sound recognition models by recording a few examples, thereby enabling personalized and fine-grained categories. ProtoSound is motivated by prior work examining sound awareness needs of DHH people and by a survey we conducted with 472 DHH participants. To evaluate ProtoSound, we characterized performance on two real-world sound datasets, showing significant improvement over state-of-the-art (e.g., +9.7% accuracy on the first dataset). We then deployed ProtoSound's end-user training and real-time recognition through a mobile application and recruited 19 hearing participants who listened to the real-world sounds and rated the accuracy across 56 locations (e.g., homes, restaurants, parks). Results show that ProtoSound personalized the model on-device in real-time and accurately learned sounds across diverse acoustic contexts. We close by discussing open challenges in personalizable sound recognition, including the need for better recording interfaces and algorithmic improvements.

Colin Lea, Zifang Huang, Dhruv Jain et al.

Voice assistants have become an essential tool for people with various disabilities because they enable complex phone- or tablet-based interactions without the need for fine-grained motor control, such as with touchscreens. However, these systems are not tuned for the unique characteristics of individuals with speech disorders, including many of those who have a motor-speech disorder, are deaf or hard of hearing, have a severe stutter, or are minimally verbal. We introduce an alternative voice-based input system which relies on sound event detection using fifteen nonverbal mouth sounds like "pop," "click," or "eh." This system was designed to work regardless of ones' speech abilities and allows full access to existing technology. In this paper, we describe the design of a dataset, model considerations for real-world deployment, and efforts towards model personalization. Our fully-supervised model achieves segment-level precision and recall of 88.6% and 88.4% on an internal dataset of 710 adults, while achieving 0.31 false positives per hour on aggressors such as speech. Five-shot personalization enables satisfactory performance in 84.5% of cases where the generic model fails.

Emma J. McDonnell, Ping Liu, Steven M. Goodman et al.

Real-time captioning is a critical accessibility tool for many d/Deaf and hard of hearing (DHH) people. While the vast majority of captioning work has focused on formal settings and technical innovations, in contrast, we investigate captioning for informal, interactive small-group conversations, which have a high degree of spontaneity and foster dynamic social interactions. This paper reports on semi-structured interviews and design probe activities we conducted with 15 DHH participants to understand their use of existing real-time captioning services and future design preferences for both in-person and remote small-group communication. We found that our participants' experiences of captioned small-group conversations are shaped by social, environmental, and technical considerations (e.g., interlocutors' pre-established relationships, the type of captioning displays available, and how far captions lag behind speech). When considering future captioning tools, participants were interested in greater feedback on non-speech elements of conversation (e.g., speaker identity, speech rate, volume) both for their personal use and to guide hearing interlocutors toward more accessible communication. We contribute a qualitative account of DHH people's real-time captioning experiences during small-group conversation and future design considerations to better support the groups being captioned, both in person and online.

Lucy Lu Wang, Kelly Mack, Emma McDonnell et al.

Accessibility research sits at the junction of several disciplines, drawing influence from HCI, disability studies, psychology, education, and more. To characterize the influences and extensions of accessibility research, we undertake a study of citation trends for accessibility and related HCI communities. We assess the diversity of venues and fields of study represented among the referenced and citing papers of 836 accessibility research papers from ASSETS and CHI, finding that though publications in computer science dominate these citation relationships, the relative proportion of citations from papers on psychology and medicine has grown over time. Though ASSETS is a more niche venue than CHI in terms of citational diversity, both conferences display standard levels of diversity among their incoming and outgoing citations when analyzed in the context of 53K papers from 13 accessibility and HCI conference venues.

Kelly Mack, Emma McDonnell, Dhruv Jain et al.

Accessibility research has grown substantially in the past few decades, yet there has been no literature review of the field. To understand current and historical trends, we created and analyzed a dataset of accessibility papers appearing at CHI and ASSETS since ASSETS' founding in 1994. We qualitatively coded areas of focus and methodological decisions for the past 10 years (2010-2019, N=506 papers), and analyzed paper counts and keywords over the full 26 years (N=836 papers). Our findings highlight areas that have received disproportionate attention and those that are underserved--for example, over 43% of papers in the past 10 years are on accessibility for blind and low vision people. We also capture common study characteristics, such as the roles of disabled and nondisabled participants as well as sample sizes (e.g., a median of 13 for participant groups with disabilities and older adults). We close by critically reflecting on gaps in the literature and offering guidance for future work in the field.

Michelle Yuan, Mozhi Zhang, Benjamin Van Durme et al.

Cross-lingual word embeddings transfer knowledge between languages: models trained on high-resource languages can predict in low-resource languages. We introduce CLIME, an interactive system to quickly refine cross-lingual word embeddings for a given classification problem. First, CLIME ranks words by their salience to the downstream task. Then, users mark similarity between keywords and their nearest neighbors in the embedding space. Finally, CLIME updates the embeddings using the annotations. We evaluate CLIME on identifying health-related text in four low-resource languages: Ilocano, Sinhalese, Tigrinya, and Uyghur. Embeddings refined by CLIME capture more nuanced word semantics and have higher test accuracy than the original embeddings. CLIME often improves accuracy faster than an active learning baseline and can be easily combined with active learning to improve results.

Leah Findlater, Steven Goodman, Yuhang Zhao et al.

Systems that augment sensory abilities are increasingly employing AI and machine learning (ML) approaches, with applications ranging from object recognition and scene description tools for blind users to sound awareness tools for d/Deaf users. However, unlike many other AI-enabled technologies, these systems provide information that is already available to non-disabled people. In this paper, we discuss unique AI fairness challenges that arise in this context, including accessibility issues with data and models, ethical implications in deciding what sensory information to convey to the user, and privacy concerns both for the primary user and for others.

Varun Kumar, Alison Smith-Renner, Leah Findlater et al.

To address the lack of comparative evaluation of Human-in-the-Loop Topic Modeling (HLTM) systems, we implement and evaluate three contrasting HLTM modeling approaches using simulation experiments. These approaches extend previously proposed frameworks, including constraints and informed prior-based methods. Users should have a sense of control in HLTM systems, so we propose a control metric to measure whether refinement operations' results match users' expectations. Informed prior-based methods provide better control than constraints, but constraints yield higher quality topics.