Statistics lecturer Dr Stéfan Janse van Rensburg has developed an innovative open-source application designed to help people with low vision read PDF documents more easily.www.mandela.ac.za reports
“The main reason was personal frustration,” Dr Janse van Rensburg, who has been diagnosed with Ocular Albinism, said. While his vision remains functional, sustained reading at normal magnification can be difficult and tiring, he explained.
“My vision is poor but still functional. I can read at normal magnification, but it’s a strain – for any sustained reading, high magnification is a practical necessity.
“Most software serves only two groups: people with normal vision or those who are clinically blind. People like me, with some usable vision, are often overlooked.”
This challenge motivated him to develop RailReader2, a desktop PDF viewer designed specifically to assist users who rely on high
Layout analysis overlay showing AI-detected document elements, such as headings, text blocks and footnotes. The AI model identifies structural components and determines the reading order within the document.
The main beneficiaries of RailReader2 are people with low vision, who rely on high magnification for reading, including students, researchers, academics and professionals.
Especially in academic environments complex PDF documents remain the dominant format for research papers, reports and textbooks.
Dr Janse van Rensburg also highlighted the potential benefits for broader communities, including ageing populations experiencing gradual vision loss and individuals in low-resource environments, who cannot afford commercial assistive technologies.
For this reason, RailReader2 has been released as a free and open-source application, available to users on both Linux and Windows platforms.
Looking ahead, he plans to expand the software’s capabilities further. One major goal is to implement continuous page rendering, to allow users to read documents as a single flowing stream, rather than page by page.
Mobile versions of the application are also considered for reading documents on smartphones and tablets.
Dr Van Rensburg’s academic work focuses on financial econometrics, time series analysis and statistical learning. Before returning to Mandela University to complete his PhD, he worked in Financial Stability at the South African Reserve Bank and lectured at Rhodes University.
His current research applies advanced statistical learning methods to renewable energy challenges, particularly solar power forecasting.
How RailReader2 works
At the centre of the application is a feature known as “rail mode”. When users zoom in beyond a certain point, the software shifts from a standard PDF viewer to a guided reading system.
Using an artificial intelligence layout model, PP-DocLayoutV3, elements, such as text blocks, headings, tables, figures and footnotes are identified and users are guided through the content line by line in logical order.
This approach helps with one of the most common problems when reading PDFs at high magnification – losing one’s place.
Dr Van Rensburg explained: “At three times magnification, a full PDF page no longer fits on the screen. RailReader2 guides users through the document structure, making it easier to follow the text.”
