Skip to main content
Have a personal or library account? Click to login
Exhibition Catalogues as Data: Sources of Structural Complexity and Steps Towards AI-Assisted Extraction Cover

Exhibition Catalogues as Data: Sources of Structural Complexity and Steps Towards AI-Assisted Extraction

Open Access
|Jul 2026

Figures & Tables

Figure 1

Sources of structural data complexity in exhibition catalogues.

Figure 2

Annotated catalogue pages illustrating data complexity in exhibition catalogues: (a) legibility issues such as faint print or overwritten text, (b) implicit reference to the artist in the previous line (‘dezelfde’), (c) ambiguous metadata formatting and structure, (d) extensive annotations including additional metadata such as lender and material information, (e) multilingual text, and (f) layout inconsistencies across catalogues. Photo RKD — Nederlands Instituut voor Kunstgeschiedenis.

Table 1

Data dictionary synthesizing the category titles, their meanings, and the reasoning behind their creation.

CATEGORYMEANING AND REASONEXAMPLE
Catalogue IDDocument title. Catalogue ID and database linker.201503502
Keyword PersonFull name of the artist. Matches category title of existing database.A. H. BAKHUIJZEN, Jr.
Artist CityCity of artist. Location marker.te ’s Hage
Artist AddressExtended address line if applicable. Additional location marker, less often specified in catalogue. Valuable for insight on artist mobility, especially for international artists who were more likely to provide full addresses.
Artist AbbreviationCatch-all category for exhibition-relevant information, such as membership and competition status. Not all exhibitions used the same abbreviations, and artists sometimes had multiple in the same entry. Created as a catch-all due to non-standardized abbreviations across exhibitions.BL
Entry NumberEntry ID. Listing number of the artwork in the catalogue.1
Free TitleTitle of the artwork. Artwork distinction and matches existing database category.Een Panorama in de omstreken van ’s Gravenhage
Additional artwork infoAnything listed about the artwork that is not part of the title, such as material, quote, ownership, or location. Created as a catch-all due to the variety of entries and inconsistencies across catalogues.unknown
AsteriskTrue/False depending on the presence of the * symbol. This symbol usually indicates whether an artwork is for sale, but can have other meanings throughout the catalogues. The meaning is only found on introductory pages.False
Amount TypeAsking price. Matches the existing database.
CurrencyCurrency in alphabetical abbreviated label, for example HFL. This format matches the existing database instead of a currency symbol or other abbreviations.
PricePrice of the artwork. Matches existing database and adds price data to artworks.
Full Entry QuoteComplete quote of the entry as formatted in the catalogue. Matches existing database and stores data without adjustments.A. H. BAKHUIJZEN, Jr., te ’s Hage. B.L. 1 Een Panorama in de omstreken van ’s Gravenhage.
Figure 3

Entry from the Tentoonstelling van schilder- en andere kunstwerken van levende meesters, held in Amsterdam by the Maatschappij Arti et Amicitiae, 1857. Photo RKD — Nederlands Instituut voor Kunstgeschiedenis.

Figure 4

Representative prompt used in the VLM workflow to extract and structure catalogue entries into predefined fields. As the model’s behavior evolved over time, the prompt required periodic refinement during the exploratory phase.

Table 2

Comparative performance of OCR+LLM and VLM pipelines across structural data challenges.

LEVELDATA CHALLENGEDEGREE OF RESOLUTIONFAILURE MODEINTERVENTION NEEDED
OCR+LLMVLM
DocumentText legibilityPartialConsistentCharacter-level noise propagates into fieldsSpot-check output and correct errors
DocumentImplicit content (e.g. “idem”, “dezelfde”)MinimalConsistentField left blank or misassignedAdjust prompts or manually propagate values
DocumentSemantic ambiguity and interpretation gapsPartialPartialClassification of values in incorrect fieldsReview and correct field assignments
DocumentPost-production annotations and marginaliaMinimalConsistentIncorrect extraction or annotation merged with entry textFilter out or verify non-original content
DocumentMultilingualityPartialConsistentLanguage switching confuses field assignmentStandardize or translate extracted values
Cross-cat.Font variationPartialConsistentCharacter forms or typographic emphasis are misread or inconsistently interpretedCorrect misread values and normalize formatting cues
Cross-cat.Layout variationPartialConsistentEntry splitting or mergingCorrect segmentation or merge entries
Cross-cat.Metadata variabilityMinimalPartialFields missing or inconsistently populatedManually align outputs to schema
Cross-cat.Language driftMinimalPartialInconsistent field assignment and lack of normalizationStandardize terms across corpus and reconcile variants
ContextualHistorical shiftsMinimalMinimalEntity variation across time prevents matchingDisambiguate entities using external sources
ContextualLegacy structuring biasMinimalMinimalConflicting field definitions and granularityReconcile fields with existing database conventions
DOI: https://doi.org/10.5334/johd.569 | Journal eISSN: 2059-481X
Language: English
Page range: 90 - 90
Submitted on: Apr 17, 2026
Accepted on: Jun 11, 2026
Published on: Jul 3, 2026
Published by: Ubiquity Press
In partnership with: Paradigm Publishing Services

© 2026 Yağmur Sarigül, Pascale Stomp, Maria Elpiniki Zafeiraki, Ava Zandieh Doulabi, Rudy Jos Beerens, Houda Lamqaddam, published by Ubiquity Press
This work is licensed under the Creative Commons Attribution 4.0 License.