Pattern recognition technology helps disabled people access books
Advances in document recognition technology have improved the lives of people with print disabilities by providing accessible documents. Twenty-five years ago, optical character recognition was practically limited to specific fonts. The creation of omnifont character recognition systems, which could read just about any machine-printed font, was the technical breakthrough that made it feasible to make a reading machine for the blind that could handle most documents. The tiny commercial market for such devices, however, made it unattractivetocommercial companies. The challenge is to take exciting new technology, such as omnifont character recognition, and get it into the hands of the people who most—but are often least able to afford—it.
We created the Arkenstone1 nonprofit social enterprise to solve this problem and deliver reading machines around the world to blind people. Arkenstone looked like a regular business with engineers, a product manager, customer support, and manufacturing. However, since it was created as a charity, the primary goal was not to make money. The goal was to help as many blind people as possible while not losing money. Arkenstone actually made money and used these profits to reinvest in more product development, providing subsidized reading machines to people on welfare, and extended the reading machine to read a dozen different languages. After 11 years, the Arkenstone reading machine business was sold to a for-profit company, and the proceeds were used to create new projects.
We have gone on to create a nonprofit umbrella organization, Benetech,2 which now has five social enterprises. Each fills a need for technology that is not lucrative enough to attract regular for-profit companies. The approach is similar to that of regular high-tech companies, with product managers assembling customer requirements and engineers creating solutions to meet those requirements. Funding comes primarily from foundations and philanthropists looking for ways to support technology for fields such as literacy, human rights, and the environment. Benetech's customers are mainly disadvantaged communities and the nonprofit organizations that serve them.
The first of these new projects is Bookshare.org, a digital library for people with disabilities. The idea is to leverage the many thousands of people who are scanning books, allowing them to share those books through an online website. This initiative is legal under US copyright law, and does not require permission or royalties as long as the only beneficiaries are people with bona fide disabilities that keep them from reading print. Bookshare.org members contribute their own books to the collection, effectively harvesting the resources of a network of blind and print-impaired people who create their own online library. After five years serving disabled Americans, Bookshare.org is expanding to other countries and languages by requesting digital permissions from authors and publishers.
One of the main technical challenges for Bookshare.org is delivering a reading experience comparable to the experience of a sighted person with a print book. While the great majority of books in the collection have been scanned by people with disabilities, their navigation through these books is limited to rough page numbering and paragraph breaks. By working with document-recognition-and-creation technologists, we hope to improve the level of text markup to assist print-disabled readers.
For example, it would be very useful to have textbooks with functional tables of contents, allowing students to navigate to chapter 7, or section 8.2, or page 379. Disabled students could then be confident that they are accessing the same material as sighted students with printed textbooks. We are working directly with publishers and authors to give us digital masters of their books to increase the stream of content that does not need to be scanned. This solution creates an additional technical challenge of converting these documents from a wide variety of digital formats to accessible formats suitable for people with disabilities. The core format in use by Bookshare.org is the ANSI/NISO Z39.86 standard3 (also known as DAISY), the XML digital talking book standard. It is easy for Bookshare.org and its users to transform books from this core format and create other accessible formats such as large print, digital Braille, text-to-speech, and MP3.
Benetech is also partnering with Adobe Systems to create improved ‘Save As XML’ capabilities for Adobe's Acrobat electronic document creation tool. The plan is to make a ‘Save As DAISY’ version of this capability that creates a DAISY document directly from a tagged PDF document. US law now requires all K–12 textbook publishers to provide schools with digital versions of textbooks in the National Instructional Materials Accessibility Standard. This new law requires textbooks to be marked up with a subset of the DAISY standard. Benetech is creating an open source tool for validating and assessing the quality of these XML documents, which should significantly improve the creation and validation of these documents.
Bookshare.org is just one example of the social enterprises we've created. Benetech is also developing solutions for people with developmental disabilities (such as Down syndrome or autism), grassroots human-rights groups, truth commissions, and environmental groups. We believe that the technology community is very interested in working on important social problems.4 The Benetech approach is a crucially important model for meeting needs when the market fails to do so.
Benetech would like to acknowledge the support of its many funders and technology partners for Bookshare.org, especially the Skoll Foundation, Omidyar Network, Schwab Foundation, the Bernard A. Newcomb Fund at Silicon Valley Community Foundation, Jenesis Foundation, the NEC Foundation of America, and Microsoft.
Jim Fruchterman is the CEO of Benetech, a leading Silicon Valley nonprofit, following founding several successful Silicon Valley high-tech companies. He concentrates on applying technology to the needs of the social sector. Fruchterman has received the MacArthur Fellowship, the Skoll Award for Social Entrepreneurship, and numerous other awards.