Interactive AI-assisted curatorial platform for tactile heritage visualization

Thesis Abstract

The rapid digitization of heritage collections has intensified the need for accessible, multi-sensory interfaces that preserve tactile affordances while expanding interpretive reach for diverse audiences. This study addresses the persistent gap in curatorial practice for tactile heritage visualization by developing an Interactive AI-assisted platform that enables researchers and visitors to explore three-dimensional replicas, haptic simulations, and multimodal annotations through an integrated AI-driven workflow. The aim is to design, implement, and evaluate a scalable platform that couples tactile data capture with intelligent curatorial reasoning to enhance material understanding and accessibility for visually impaired and sighted users alike. Specific objectives are (i) to operationalize a modular architecture integrating 3D tactile models, haptic feedback, and textual/visual metadata; (ii) to develop AI components for artifact classification, anomaly detection in tactile scans, and personalized curation based on user profiles; (iii) to assess the platform’s impact on learning outcomes, engagement, and accessibility satisfaction; (iv) to investigate ethical and provenance considerations in AI-assisted tactile representation; and (v) to formulate guidelines for sustainable deployment in museum and archival settings. The methodology adopts a pragmatic mixed-methods design. The population consists of 40 curators and 60 museum visitors drawn from three regional museums with active tactile-heritage programs. A purposive sampling approach targets professionals with experience in tactile displays and AI-enabled exhibits, complemented by a random stratified sample of visitors to capture diverse user experiences. Data collection instruments include (a) a structured survey instrument with validated scales for perceived usefulness, perceived ease of use, and accessibility satisfaction (n=100); (b) semi-structured interviews with 18 curators to explore operational challenges, ethical considerations, and curatorial authority; (c) task-based usability testing involving 24 participants performing predefined interpretive tasks using the platform; and (d) log data capturing interaction sequences, dwell times, and haptic modality usage. The primary data analysis employs descriptive statistics and regression analysis to identify factors predicting user engagement and learning gains, thematic analysis of interview transcripts guided by Braun and Clarke’s approach for qualitative insight, and ANOVA to compare outcomes across user groups. The study also integrates a model-based evaluation of the AI components, including precision and recall for artifact classification, and user-model alignment metrics for personalization. A conceptual framework grounded in Activity Theory and the Extended UTAUT model guides interpretation of results, with particular attention to-mediated action, community of practice, and technology acceptance in tactile curatorial contexts. The interactive platform comprises (i) a tactile-visual corpus representing high-fidelity replicas and haptic-enabled surfaces; (ii) AI modules for 3D feature extraction, texture synthesis, and contextual recommendation; (iii) a curatorial dashboard enabling provenance tracking, annotation, and versioning; and (iv) accessibility adapters including audio narration and Braille/large-print overlays. The study anticipates that the platform will yield statistically significant improvements in learning outcomes (p<0.05), higher engagement scores among visually impaired participants, and increased efficiency in curatorial workflows. Potential findings include enhanced accuracy in tactile interpretation due to AI-assisted feature highlighting, greater inclusivity reflected in user satisfaction across demographic groups, and identifiable trade-offs between automation and expert curatorial authority requiring governance protocols. The anticipated contribution to knowledge includes a novel, scalable model for AI-assisted tactile heritage visualization that integrates curatorial practice with inclusive design, an empirical evaluation of AI-driven tactile interpretation, and a set of guidelines for ethical AI deployment in cultural heritage contexts. The study concludes that AI-enabled tactile visualization can expand access to museum collections without compromising interpretive rigor if accompanied by transparent provenance, human-in-the-loop governance, and ongoing portability testing. Recommendations address standardization of tactile data formats, iterative user-centered design cycles, and policy development for accessibility funding, metadata interoperability, and cross-institutional collaboration to sustain repertoires of tactile heritage.