Accessible On-line Graphics
Anuradha Madugalla · 2015 · Proceedings of the 12th International Web for All Conference (W4A) · doi:10.1145/2578726.2746672
Summary
This doctoral consortium paper proposes semi-automated methods for transcribing online graphics — particularly floor plans — into accessible formats for blind users. The research is motivated by the fundamental problem that screen readers and voice synthesisers can only process text, leaving the vast amount of graphical content on the web inaccessible to blind users. Current methods for creating accessible graphics are inadequate: annotation at the development stage fails due to deadline pressure and annotation not being a client priority, while manual transcription by expert transcribers using tools like CorelDRAW is prohibitively expensive and slow — Australia's State-wide Vision Resource Centre estimates that transcribing a single graphics-heavy textbook costs AUD 120,000 and takes several months. The research focuses specifically on floor plans as an underexplored but important type of information graphic. While accessible maps have advanced significantly using sonification, speech, and vibration, floor plans have received comparatively little attention despite their value for enabling blind users to independently navigate buildings like airports, shopping centres, and apartments.
Key findings
The research proposes three progressive aims: first, a transcription tool for raster-based floor plans using image processing techniques; second, a transcription tool for SVG-based floor plans, motivated by SVG's scalability advantages (raster graphics pixelate when zoomed, which is especially problematic since blind users of tactile displays need frequent zoom level changes due to lower finger resolution compared to eyes); and third, a general transcription tool for any SVG graphic. All three tools would output to GraVVITAS, an iPad app developed in collaboration with Vision Australia at Monash University that enables blind users to freely interact with graphics through sonification and voice feedback on a touchscreen. The paper reviews prior work on SVG accessibility: Rotard et al. explored SVG bar charts but required manual intervention and predefined formats; the GATE project automated SVG transcription but required pre-annotated graphics; and GraSSML captured structure and semantics of arbitrary SVG diagrams but was heavily dependent on domain knowledge. The proposed research aims to handle arbitrary SVG graphics from any domain without requiring predefined formats — identified as an uncharted area. Floor plan complexity is emphasised: even house plans contain rooms, doors (of different types affecting how they open), windows, fixed furniture (benches, sinks), and movable furniture (chairs, tables), all of which must be conveyed for confident navigation.
Relevance
This paper addresses a persistent gap in web accessibility: while text alternatives for images are well-understood, complex information graphics like floor plans, charts, and diagrams remain largely inaccessible to blind users. The economic argument is compelling — manual transcription is so expensive that it cannot scale to meet the volume of online graphics. The semi-automated approach using image processing with human verification offers a pragmatic middle ground. The focus on floor plans fills a practical need: being able to preview a building's layout before visiting could significantly increase independence for blind people in unfamiliar environments. The argument for SVG over raster formats as the future of accessible web graphics is well-founded — SVG's scalability, structured data, and semantic potential make it inherently more accessible than pixel-based formats. However, this is a proposal paper with no completed tools or user evaluation. The challenge of extracting semantic meaning from arbitrary graphics (particularly raster images) without domain knowledge remains a very hard problem. The reliance on an iPad-specific app (GraVVITAS) also limits platform reach.
Tags: blindness · tactile graphics · accessible graphics · floor plans · SVG · image processing · sonification · indoor navigation · graphic transcription