With the continuous development of society and technology, accessibility for minority groups, particularly visually impaired persons (VIPs), is gaining more attention. This paper explores ways to enhance VIP reading efficiency and experience through the design of a tailored automatic text summarization tool, contributing to the European Accessibility Act's objective to improve accessible products and services. In collaboration between Delft University of Technology and Koninklijke Bibliotheek, this project addresses the specific challenges VIPs face when accessing information in both physical and digital library environments.
This research highlights the gap between sighted readers and VIPs in skimming reading strategies, a critical aspect of information retrieval. Skimming, defined as quickly identifying the main ideas of a text, is typically a faster process for sighted readers who can use spatial cues. For VIPs, reading is linear and auditory, increasing their cognitive load and reducing comprehension. Through a detailed analysis of VIP needs and existing skim-reading tools, such as Skimmer, this paper narrows its design scope to address the identified gaps, particularly in two-dimensional navigation.
The primary goal of this research is to develop an AI-driven summarization tool that enhances the skim-reading experience for VIPs by supporting non-visual implementations of macro-structures and enabling non-linear exploration. By tailoring summaries to VIP preferences, the tool allows users to quickly access relevant sections of text, improving both reading speed and overall comprehension. The design employs inclusive, iterative, and user-centered methods, ensuring the final product meets the specific needs of VIPs. User testing and feedback are integral to the development process, refining the tool to enhance usability and accessibility.
The project also contributes to the theoretical understanding of VIPs’ cognitive processes during skim reading. It reveals the importance of designing summaries that are intuitive and customizable to meet diverse needs. By integrating generative AI tools, the tool can produce personalized summaries, progressively refining them to align with user interests. The final deliverable is a web-based application that provides VIPs with efficient and satisfying reading experiences, closing the gap between sighted and visually impaired readers.
This research advances assistive technology by addressing the lack of suitable reading tools for VIPs, particularly in non-academic, everyday reading contexts. The project fills market gaps by offering an accessible solution for digital reading materials that can cater to the unique challenges faced by VIPs. Ultimately, the tool enhances VIPs' cognitive abilities, reduces their cognitive burden, and fosters a more engaging reading experience, contributing to increased reading efficiency and self-efficacy.

With the continuous development of society and technology, accessibility for minority groups, particularly visually impaired persons (VIPs), is gaining more attention. This paper explores ways to enhance VIP reading efficiency and experience through the design of a tailored automatic text summarization tool, contributing to the European Accessibility Act's objective to improve accessible products and services. In collaboration between Delft University of Technology and Koninklijke Bibliotheek, this project addresses the specific challenges VIPs face when accessing information in both physical and digital library environments.
This research highlights the gap between sighted readers and VIPs in skimming reading strategies, a critical aspect of information retrieval. Skimming, defined as quickly identifying the main ideas of a text, is typically a faster process for sighted readers who can use spatial cues. For VIPs, reading is linear and auditory, increasing their cognitive load and reducing comprehension. Through a detailed analysis of VIP needs and existing skim-reading tools, such as Skimmer, this paper narrows its design scope to address the identified gaps, particularly in two-dimensional navigation.
The primary goal of this research is to develop an AI-driven summarization tool that enhances the skim-reading experience for VIPs by supporting non-visual implementations of macro-structures and enabling non-linear exploration. By tailoring summaries to VIP preferences, the tool allows users to quickly access relevant sections of text, improving both reading speed and overall comprehension. The design employs inclusive, iterative, and user-centered methods, ensuring the final product meets the specific needs of VIPs. User testing and feedback are integral to the development process, refining the tool to enhance usability and accessibility.
The project also contributes to the theoretical understanding of VIPs’ cognitive processes during skim reading. It reveals the importance of designing summaries that are intuitive and customizable to meet diverse needs. By integrating generative AI tools, the tool can produce personalized summaries, progressively refining them to align with user interests. The final deliverable is a web-based application that provides VIPs with efficient and satisfying reading experiences, closing the gap between sighted and visually impaired readers.
This research advances assistive technology by addressing the lack of suitable reading tools for VIPs, particularly in non-academic, everyday reading contexts. The project fills market gaps by offering an accessible solution for digital reading materials that can cater to the unique challenges faced by VIPs. Ultimately, the tool enhances VIPs' cognitive abilities, reduces their cognitive burden, and fosters a more engaging reading experience, contributing to increased reading efficiency and self-efficacy.