The 15th Web for All Conference Lyon, France 2018

Session 10: Evaluation and Testing

Automatic Role Detection of Visual Elements of Web Pages for Automatic Accessibility Evaluation




Carlos Duarte, Ana Salvado, Luis Carrico, LASIGE, Faculdade de Ciências da Universidade de Lisboa; M Elgin Akpinar, Yeliz Yesilada, Middle East Technical University




Understanding what role a visual element plays in a web page can have a significant impact in the accessibility of that page. From an accessibility conformance testing perspective, several WCAG techniques require role identification of some visual elements during the testing process. This has prevented automatic accessibility evaluation tools from processing these techniques. A technique for automatic role detection of visual elements of web pages was recently developed as a complement to a web page segmentation process. This paper explores this role identification technique in the context of automatic web accessibility evaluation. The technique is integrated with an automated evaluation tool, and demonstrated to successfully evaluate some WCAG techniques that rely on the identification of visual elements, in particular menu and list elements.




Notes:

Semantic Content Analysis Supporting Web Accessibility Evaluation




Carlos Duarte, Ines Matos, Luis Carrico, LASIGE, Faculdade de Ciências da Universidade de Lisboa




Despite its many advantages, automated web accessibility evaluation’s main limitation is still its inability to assess rules and techniques that require a semantic understanding of the web content. Today, machine learning solutions are available that can interpret different media content with a reasonable degree of confidence. These solutions have an untapped potential to increase the accessibility of web pages. This potential extends to the evaluation of web pages also. This paper proposes an algorithm to automatically rate the similarity between a content and its textual description in a web accessibility evaluation context. The validity of the algorithm is demonstrated by comparing its ratings of descriptions of images obtained from their alt texts with human ratings of the same descriptions.




Notes:

Web Accessibility Evaluation in a Crowdsourcing-Based System with Expertise-Based Decision Strategy




Shuyi Song, Bu Jiajun, Ye Wang, Zhi Yu, Andreas Artmeier, Can Wang, Zhejiang University; Lianjun Dai, China Disabled Persons' Federation Information Center




The rising awareness of accessibility increases the demand for Web accessibility evaluation projects to verify the implementation of Web accessibility guidelines and identify accessibility barriers in websites. However, the complexity of accessibility evaluation tasks and the lack of experts limits their scope and reduces their significance. Due to this complexity, they could not directly rely on a technique called crowdsourcing, which made great contributions in many fields by dividing a problem into many tedious micro-tasks and solving tasks in parallel. Addressing this issue, we develop a new crowdsourcing-based Web accessibility evaluation system with two novel decision strategies, golden set strategy and time-based golden set strategy. These strategies enable the generation of task results with high accuracy synthesized from micro-tasks solved by workers with heterogeneous expertise. An accessibility evaluation of 98 websites by 55 workers with varying experience verifies that our system can complete the evaluation in half the time with a 7.21% improvement on accuracy than the current approach.




Notes:

Reliability Aware Web Accessibility Experience Metric




Shuyi Song, Jiajun Bu, Chengchao Shen, Andreas Artmeier, Zhi Yu, Zhejiang University; Qin Zhou, China Disabled Persons' Federation Information Center




Web accessibility metrics measure the accessibility levels of websites, especially for people with disabilities. Although many metrics with different motivations have been proposed in recent years, current metrics are limited in their applicability when considering user experience. In this study, we propose Reliability Aware Web Accessibility Experience Metric (RA-WAEM), a novel Web accessibility metric which considers the user experience of people with disabilities by their provided pairwise comparisons of user experience between different websites and their heterogeneous reliability in objectively accessing the severity of accessibility barriers. Since determining the reliability of users is difficult and time-consuming, we develop an optimization algorithm based on Expectation Maximization to derive the parameters of RA-WAEM. Furthermore, we present a complete Web Accessibility evaluation framework, which we facilitate to conduct an extensive accessibility study on 46 websites with 323,098 Web pages and collect the user experience of 122 people on these websites. An evaluation on this dataset shows that RA-WAEM outperforms state of the art Web accessibility metrics in reflecting the user experience of people with disabilities.

Notes:

Session 11: Doctoral Consortium Presentations - 2 of 2

Arabic Accessibility Guidelines – Understanding and Use by Web Developers in Kuwait

Muhammad Saleem, Edith Cowan University, Western Australia

The aim of this research is to develop and implement Arabic accessibility resources for developers, web content managers and designers. The Arabic guidelines will not only assist Arabian developers and designers for a deep understanding of accessibility features, but also to apply these criteria on their Arabic websites in order to make them accessible to everyone including people with disabilities. The Arabic web accessibility guidelines will be designed to be reachable to all developers and designers in the Middle East including Kuwait.

Using Ontologies as a Foundation for Web Accessibility Tools

Jens Pelzetter, University of Bremen, Germany

Creating web sites has become quite a complex task. One of most important aspects of a modern web site is accessibility. However, despite extensive standards many web sites have accessibility issues. This paper presents a new approach for creating tools to improve the accessibility of web sites using ontologies.

Notes:

Session 12: Assistive Technology

Exploring Aural Navigation by Screenless Access

Mikaylah Gross, Joe Dara, Christopher Meyer, Davide Bolchin, IU School of Informatics and Computing – Indianapolis




When the Blind and Visually Impaired (BVI) navigate the mobile web, they have to hold a phone in their hands at all times. Such continuous, two-handed interaction on a small screen hampers the user’s ability to keep hands free to control aiding devices (e.g., cane) or touch objects nearby, especially on-the-go. In this paper, we introduce screenless access: a browsing approach that enables BVI to interact touch-free with aural navigation architectures using one-handed, in-air gestures recognized by an off-the-shelf armband. In a study with ten BVI participants, we observed fast learnability of the gestures, good aural navigation performance, user’s conceptual fit with a screen-free paradigm, and low levels of cognitive load. Our findings uncovered distinct navigation styles that participants used when approaching screenless browsing and a range of unprompted adaptations of touch-free gestures and arm-body postures that were effectively enacted to appropriate the technology. User feedback revealed insights into the potential and limitations of screenless navigation to support convenience in traveling, work contexts and privacy-preserving scenarios, as well as concerns for gestures that may become socially conspicuous.




Notes:

Multi-view Mouth Renderization for Assisting Lip-reading

Andrea Britto Mattos, Dario Augusto Borges Oliveira, IBM Research

Previous work demonstrated that people who rely on lip-reading often prefer to have a frontal view of their interlocutor, but sometimes a profile view may display certain lip gestures more noticeably. This work refers to an assistive tool that receives an unconstrained video of a speaker, captured at an arbitrary view, and not only locates the mouth region but also displays augmented versions of the lips in the frontal and profile views. This is made using a deep Generative Adversarial Network (GAN) trained on several pairs of images. In the training set, each pair contains a mouth picture taken at a random angle and the corresponding picture (i.e., relative to the same mouth shape, person, and lighting condition) taken at the frontal view. In the test phase, the network is able to receive an unseen mouth image taken at an arbitrary view and map it to the frontal position -- the profile views are rendered accordingly. Because building a large-scale pairwise dataset is time consuming, we use realistic synthetic 3D models for training, and videos of real subjects as input for testing. Our approach is speaker-independent, language-independent and our results demonstrate that the GAN is able to produce visually compelling results that may assist people with hearing impairment.

Notes:

Accessify: An ML Powered Application to Automatically Provide Image Descriptions on Website

Shivam Singh, Anurag Bhandari, Nishith Parak, Accenture

Web Accessibility is one of the major concerns for rapidly growing internet community. Accessing online content is crucial for many people as they engage in variety of tasks every day. The content can be of any form like audio, video, images and text. A major portion of website screen generally contains images and there are set standards for providing information about these images while creating the image component. Yet, most of the website available online do not strictly follow the accessibility guidelines and as a result, either the website contains no description of the image or the description provided is not sufficient enough to convey the meaning to visually impaired people who use screen readers for accessing the web content. In this paper, we propose a solution by providing alternate image descriptions to all images on any website with help of Machine Learning in a complete unobtrusive environment. The application does not require any setup and can be used on static as well as dynamic websites.

Notes:

Session 13: Motor Impairments

Vocal Programming for People with Upper-Body Motor Impairments

Lucas Rosenblatt, Brown University; Patrick Carrington, Jeffrey Bigham, Carnegie Mellon University;Kotaro Hara, Singapore Management University

Programming heavily relies on entering text using traditional QWERTY keyboards, which poses challenges for people with limited upper-body movement. Developing tools using a publicly available speech recognition API could provide a basis for keyboard free programming. In this paper, we describe our efforts in design, development, and evaluation of a voice-based IDE to support people with limited dexterity. We report on a formative Wizard of Oz (WOz) based design process to gain an understanding of how people would use and what they expect from a speech-based programming environment. Informed by the findings from the WOz, we developed VocalIDE, a prototype speech-based IDE with features such as Context Color Editing that facilitates vocal programming. Finally, we evaluate the utility of VocalIDE with 8 participants who have upper limb motor impairments. The study showed that VocalIDE significantly improves the participants’ ability to make navigational edits and select text while programming.




Notes:

Designing an Adaptive Web Navigation Interface for Users with Variable Pointing Performance

Aqueasha Martin-Hammond, Purdue University Indianapolis; Foad Hamidi, Tajas Bhalerao, Catherine Hornback, Amy Hurst, University of Maryland, Baltimore County; Abdullah Ali, University of Washington; Casey Means, Rhodes College




Many online services and products require users to point and interact with user interface elements. For individuals who experience variable pointing ability due to physical impairments, environmental issues or age, using an input device (e.g., a computer mouse) to select elements on a website can be difficult. Adaptive user interfaces dynamically change their functionality in response to user behavior. They can support individuals with variable pointing abilities by 1) adapting dynamically to make element selection easier when a user is experiencing pointing difficulties, and 2) informing users about these pointing errors. While adaptive interfaces are increasingly prevalent on the Web, little is known about the preferences and expectations of users with variable pointing abilities and how to design systems that dynamically support them given these preferences.

We conducted an investigation with 27 individuals who intermittently experience pointing problems to inform the design of an adaptive interface for web navigation. We used a functional high-fidelity prototype as a probe to gather information about user preferences and expectations. Our participants expected the system to recognize and integrate their preferences for how pointing tasks were carried out, preferred to receive information about system functionality and wanted to be in control of the interaction. We used findings from the study to inform the design of an adaptive Web navigation interface, PINATA, that tracks user pointing performance over time and provides dynamic notifications and assistance tailored to their specifications. Our work contributes to a better understanding of users’ preferences and expectations of the design of an adaptive pointing system.




Notes:


Collaborative Accessibility Assessments by Senior Citizens Using Smartphone Application ReAcTS (Real-world Accessibility Transaction System)

Takahiro Miura, Ken-ichiro Yabu, Institute of Gerontology (IOG), The University of Tokyo




Rapid improvements in real-world accessibility conditions have enabled greater mobility of people with disabilities as well as senior citizens.However, while the accessibility infor- mation of downtown areas and sightseeing locations is fre- quently updated, the most recent information on accessibility conditions in areas such as suburbs is difficult to obtain quickly because of infrequent updates of this type of information along with local information upgrades. On the other hand, healthy seniors aged over 65 years are known to engage in volunteer activities in developed countries with a high population of the elderly. In this regard, collaborative work on the accessibility assessments of places unexplored by these senior citizens can enable not only the efficient data collection on local areas but also lead to increase in their awareness and motivation to participate in social activities. In this study, we propose an accessibility-sharing application for volunteers including se- niors and youngsters along with a scheme to efficiently record real-world conditions. In particular, we develop a smartphone application named Real-world Accessibility Transaction Sys- tem (ReAcTS), and we conduct events to share accessibility information in cooperation with volunteer groups mainly composed of senior citizens.

Notes:

Awards and Closing

W4A Sponsors & Supporters

Sponsors: · IBM · The Paciello Group · Google · Intuit
Supporters:
· WebKeyIt · Capti · OpenConf
In Cooperation with:
· TheWeb Conference · ACM · sigweb

Conference Organization

General Chairs

· Chris Bailey – UXPA UK, Enabling Insights Ltd.

· Jeff Bigham – Carnegie Mellon University, USA

Program Chairs

· Volker Sorge – University of Birmingham, UK

· Elaine Pearson – Teesside University, UK

Challenge Chairs

· Silvia Mirri – University of Bologna, Italy

· David Sloan – The Paciello Group, UK.

Doctoral Consortium

· Dragan Ahmetovic – Carnegie Mellon University, USA

· Giorgio Brajnik – University of Udine, Italy

IBM Awards Chairs

· Chieko Asakawa -IBM Research, Japan

· Yevgen Borodin -Stony Brook University, USA

Student Awards Chairs

· Chieko Asakawa – IBM Research, Japan

· Yevgen Borodin – Stony Brook University, USA.

Steering Committee

· Vivienne Conway (2017-) Edith Cowan University (GC Support)

· Luz Rello (2017- ) Carnegie Mellon University (Media)

· Greg Gay (2016- ) Ryerson University, Canada (Website)

· Yevgen Borodin (2014-) Stony Brook University, USA. (Publicity chair)

· Markel Vigo (2012-) University of Manchester, UK. (Finance)

· Yeliz Yesilada (Founder/2003-) Middle East Technical University (METU), Northern Cyprus. (Sponsorship)

· Simon Harper (Founder/2003-) University of Manchester, UK. (WWW Liaison)

Advisory Committee

· Tiago Guerreiro (2016-) University of Lisbon, Portugal

· Luis Carriço (2015-) University of Lisbon, Portugal

· Silvia Mirri (2015-) University of Bologna, Italy

· Jeff Bigham (2014-) Carnegie Mellon University, USA

· Giorgio Brajnik (2013-) University of Udine, Italy

· Leo Ferres (2011-) Universidad de Concepción, Chile

· Hironobu Takagi (2010-) IBM Research, Japan

· Chieko Asakawa (2010-) IBM Research, Japan

· David Sloan (2009-) The Paciello Group, UK

Program Committee

· Julio Abascal – University of the Basque Country

· Shadi Abou-Zahra – W3C

· Eyhab Al-Masri – Wilfrid Laurier University

· Vikas Ashok – Stony Brook University

· Jennison Asuncion – LinkedIn

· Giorgio Brajnik – University of Udine

· Rocio Calvo – Universidad Carlos III de Madrid

· Olga De Troyer – Vrije Universiteit Brussel

· Ted Drake – INUIT

· Carlos Duarte – University of Lisbon

· Sukru Eraslan – Middle East Technical University

· Greg Gay – Ryerson University

· Voula Gkatzidou – Brunel University London

· Steve Green – Reedscar Limited

· Simon Harper – University of Manchester

· Andy Heath – Independent Consultant

· Amy Hurst – University of Maryland

· Hernisa Kacorri – University of Maryland

· Shaun Kane – University of Colorado Boulder

· Barbara Leporini – CNR-ISTI

· Aqueasha Martin – University of Maryland Baltimore County

· Kyle Montague – Newcastle University

· Hugo Nicolau – University of Lisbon

· Mike Paciello – The Paciello Group

· Fabio Paterno – CNR-ISTI

· Enrico Pontelli – New Mexico State University

· I.V. Ramakrishnan – Stony Brook University

· Paola Salomoni – University of Bologna

· David Sloan – The Paciello Group

· Neil Soiffer – Talking Cat Software

· Andrii Soviak – Stony Brook University

· Phil Strain – Spotify

· Markel Vigo – University of Manchester

· Konstantinos Votis – Centre for Research and Technology Hellas

· Yeliz Yesilada – Middle East Technical University

· Yu Zhong – Google