Accepted Travel Scholarship Participants

TS-Jarjue Revisiting Blind Photography in the Context of Teachable Object Recognizers
Ebrima Jarjue, University of Maryland College Park
Transcript of Audio File

The title of the research poster is Captioning Practices in Deaf Vlogs on YouTube

The authors are Julia Kerns, Elizabeth Beers, Hernisa Kacorri from the University of Maryland, College Park, MD.

This work aims to provide a better understanding of current practices in captioning of Deaf Vlogs in a popular medium such as Youtube. Our observations can have implications for future work in building sign language datasets scrapped from the web. Moreover, they can inform the design of vlogging interfaces by considering the Deaf community’s ownership over signed content in the presence of future auto-translation features.

To collect our data, we used the keyword deaf vlog which we then used to code the top 100 videos however videos were not restricted to a specific sign language. For a video to be included, it had to include sign language content created by Deaf people.

We used several methods in characterizing the videos: 1. Type of Sign Language, 2. types of captioning practices which are none, user-generated closed captions, auto-generated closed captions, open captions, audio, and transcripts, 3. thematic category listed by YouTube, 4. Intended audience, 5. Number of views

We concluded that over 42 of the videos were intended for a Deaf audience. The distribution of the number of views by captioning practices was higher throughout the videos containing closed captioning and audio. These results provide us with the understanding that there is more work to be done on creating datasets to analyze the various types of captioning and various types of sign language used in these videos. This information will provide us with a foundational start to begin indexing Deaf Vlogs for future technologies. Thank you.

Transcript of Audio File

IMPACT OF ASSISTIVE TECHNOLOGIES IN SUPPORTING PEOPLE WITH DEMENTIA

Dr. Ikram Asghar

Introduction

Due to ageing, global dementia population has reached over 50 million. The use of assistive technologies has the potential to help them.

However, these technologies have high abandonment rate due to challenges like user friendliness, social isolation and participatory design.

Aims and Objectives

This study aims at empirically evaluating the impacts of assistive technology to support the people with dementia. The research objectives are:

Investigate existing assistive technologies

Identify important factors

Elicit requirements

Design and development new technology

Impact experiments

Methodology

This study used mixed research methodology approach. Secondary research methods including literature review and systematic mapping.

Primary research methods including interviews, questionnaires and case studies.

Proposed Solution

Developed an Assistive Brotherhood Community Software Application

Conducted eight case studies involving 8 dementia people and 40 volunteers

People with dementia got their needed items quickly and in friendlier manner through the help of volunteers

The involvement of the caregivers for medication, meals, and prayers reduced significantly

The painting function evoked memories and improved motivation

The news and weather functions kept updated

The travel tutor guided in safe travel outside home and made sure that they get back independently

Research Contributions and Impacts

Largest study globally to involve the 347 people with dementia

First usability evaluation study in South Asia

National dementia strategy help for developing countries

First day care centre at Pakistan

Six journal and three conference publications

Future Work

Planning to implement Assistive Brotherhood Communities in Western Countries.