Beyond Beautiful: Embroidering Legible and Expressive Tactile Graphics
Margaret Ellen Seehorn, Claris Winston, Bo Liu, Gene S-H Kim, Emily White, Nupur Gorkar, Kate S Glazko, Aashaka Desai, Jerry Cao, Megan Hofmann, Jennifer Mankoff · 2025 · ASSETS 2025: 27th International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/3663547.3746336
Summary
This paper explores machine embroidery as a medium for producing tactile graphics—tangible representations of visual information used by blind and visually impaired (BVI) individuals. Current methods for producing tactile graphics, primarily embossing and swell paper printing, have significant limitations: embossed materials degrade over time, swell paper offers limited texture variety (primarily raised dots and lines), and the specialized equipment is expensive and often only available in institutional settings. Embroidery offers advantages including durability, a broad range of distinguishable stitch textures, material softness suitable for active use, and accessibility of the craft to a wide community of makers. The research follows a multi-phase approach. First, the team manually designed embroidered tactile graphics in two domains—a children's book ("Where the Wild Things Are") and scientific diagrams (bar chart, heart diagram)—to understand design requirements. These explorations revealed that visually distinct stitch patterns do not necessarily produce tactilely distinct textures, highlighting the need for systematic texture categorization. Second, the researchers compiled 55 embroidery textures from the Embrilliance software and systematically categorized them across seven measurable properties: number of stitches, weight, height, whitespace, direction, motif category, and stitch complexity. Third, they developed an optimization algorithm using stochastic hill-climbing to automatically assign maximally distinguishable textures to regions of tactile graphics, optimizing for both neighboring contrast and overall contrast across the graphic. Fourth, they built an end-to-end pipeline from SVG input through texture assignment, machine embroidery printing, Braille integration (printed on the reverse side of the fabric), and post-processing. The pipeline uses a Janome embroidery machine, satin fabric backing, and the Embrilliance software.
Key findings
A user study with six blind participants evaluated four embroidered tactile graphics (paramecium diagram, geological layers, campus map, floor plan). Participants answered 70 out of 78 comprehension questions correctly (89.74%), demonstrating that embroidered tactile graphics present information accurately and comprehensibly. Average ratings on a 7-point scale were: supportiveness 6.13, comprehensibility 6.08, and confidence 6.19. Key design insights emerged: (1) textural richness of embroidery must be measured and verified, not assumed from visual appearance—visually different textures can feel similar tactilely; (2) spacing and whitespace between regions significantly aids legibility, with participants preferring fewer, more distinct textures over many similar ones; (3) holistic graphic design matters as much as individual texture selection—the backing material, border treatments, and overall composition affect comprehension; (4) embroidered Braille labels, while initially challenging (softer and more spaced than traditional embossed Braille), became more readable with practice; (5) participants found embroidered graphics less "monotonous" than embossed alternatives and appreciated the broader range of expressive textures. The optimization algorithm successfully assigned distinct textures that maximized both neighboring and overall contrast. Production takes 2-4 hours per graphic on a machine costing approximately \,000.
Relevance
This research opens a promising new production method for tactile graphics that addresses key limitations of current approaches. For accessibility practitioners and educators, embroidery offers practical advantages: durability for repeated use (unlike degradable swell paper), a wider range of distinguishable textures for representing complex diagrams, and the potential to engage a broader community of makers (embroidery is a widespread craft skill). The systematic categorization of tactile texture properties—number of stitches, weight, height, whitespace, direction, motif, and complexity—provides a framework that could inform any tactile graphic design, not just embroidered ones. The finding that visual distinctiveness does not predict tactile distinctiveness is a crucial insight for anyone designing tactile materials. The end-to-end pipeline, while still requiring some manual steps, represents progress toward making tactile graphic production more accessible. Limitations include the small participant sample, the cost of embroidery machines, the current reliance on manual SVG preparation, and the inaccessibility of some embroidery software to BVI users themselves.
Tags: tactile graphics · machine embroidery · blind and visually impaired · fabrication · braille · texture design · optimization · accessibility
Standards referenced: Braille Authority of North America Tactile Graphics Guidelines · Unified English Braille