Guiding Novice Web Workers in Making Image Descriptions Using Templates

Abstract

This article compares two methods of employing novice Web workers to author descriptions of science, technology, engineering, and mathematics images to make them accessible to individuals with visual and print-reading disabilities. The goal is to identify methods of creating image descriptions that are inexpensive, effective, and follow established accessibility guidelines. The first method explicitly presented the guidelines to the worker, then the worker constructed the image description in an empty text box and table. The second method queried the worker for image information and then used responses to construct a templatebased description according to established guidelines. The descriptions generated through queried image description (QID) were more likely to include information on the image category, title, caption, and units. They were also more similar to one another, based on Jaccard distances of q-grams, indicating that their word usage and structure were more standardized. Last, the workers preferred describing images using QID and found the task easier. Therefore, explicit instruction on image-description guidelines is not sufficient to produce quality image descriptions when using novice Web workers. Instead, it is better to provide information about images, then generate descriptions from responses using templates.

Accessible Visualization via Natural Language Descriptions: A Four-Level Model of Semantic Content

“…While these approaches are well-intentioned, in aiming to address the engineering problem of how to automatically generate natural language captions for charts, they have largely sidestepped the complementary (and prior) question: which semantic content should be generated to begin with? Some captions may be more or less descriptive than others, and different readers may receive different semantic content as more or less useful, depending on their levels of data literacy, domain-expertise, and/or visual perceptual ability [69,71,72]. To help orient work on automatic visualization captioning, our four-level model of semantic content offers a means of asking and answering these more human-centric questions.…”

SlideImages: A Dataset for Educational Image Classification

“…We intend to use document figure classification as a first step in automatic educational image summarization applications. A similar idea is followed by Morash et al [10], who built one template for each type of image, then manually classified images and filled out the templates, and suggested automating the steps of that process. Moraes et al [11] mentioned the same idea for their SIGHT (Summarizing Information GrapHics Textually) system.…”

"I Hope This Is Helpful"

“…Such close-ended approaches might potentially even go so far as to provide a fixed checklist instead of a free-text box. Alternatively, templates could be created that, for example, identify what types of content to insert when constructing a sentence [77,107]. Such approaches "prime" workers towards certain work outcomes by giving them examples.…”