Crowdsourcing descriptions of images on museum websites for visually impaired visitors: developing guidelines and examples of good practicePaper
Helen Petrie, UK
Many museums are now providing thousands of images of items from their collections on their websites. Members of the public, researchers and students use these images for many purposes. However, people with visual impairments, who may be interested in using information from such images, are usually unable to do so, even though they can access other information on the website using text-to-speech technology. Providing descriptions of all the images on a museum website would be a very costly undertaking, so Museum Victoria has been exploring the use of crowdsourcing to solve this issue, asking sighted members of the public to describe images for visually impaired people.
One important aspect of such an initiative is how to guide members of the public in describing images appropriately for visually impaired people. The Web Content Accessibility Guidelines (WCAG) developed by the World Wide Web Consortium advocate describing images on the Web for visually impaired Web users, but provide no information about what the content of such descriptions should actually be. Other information about how to describe images on the Web for visually impaired people is very scarce, and may not be applicable to museum images.
Therefore we have undertaken a series of three studies to develop a set of guidelines and examples of good practice on image description for museum websites, working closely with visually impaired people and experts in visual impairment. Study 1 involved interviews with 17 visually impaired people (aged 29 – 80, 7 women and 10 men) about what kinds of information they would like about images on museum websites. Seven images from the Museum Victoria online collection were described to them and these provided concrete examples on which to base discussions. Analysis of these interviews formed the basis for the first version of a set of guidelines and examples of good practice for image description. One key finding from this study was that the visually impaired participants would like a “short” description to give them a brief overview of the image and a “long” description, for more detail. However, it was not clear how short a “short” description should be or how long a “long” description should be.
In Study 2, 11 members of the public with no particular knowledge of visual impairment were asked to read the guidelines and examples of good practice and then provide short and long descriptions of four of the images from the Museum Victoria website (from Study 1). The participants were also asked to comment on the usefulness of the guidelines and examples. The guidelines and examples were also given to six experts in visual impairment for their comments and feedback.
For Study 3, the descriptions created in Study 2 were analysed and a series of typical “short” and “long” descriptions of different lengths of the four images were created. These were read to eight of the visually impaired participants from Study 1 and discussed with them. Analysis of Studies 2 and 3 allowed the refinement of the guidelines and examples of good practice.