Paris Archives and Museums Tackle Duplicate Image Crisis After Audit Flags Thousands of Redundant Files

Paris's major cultural institutions are scrambling this week to address a growing digital housekeeping crisis after an internal audit circulated among Île-de-France heritage bodies identified tens of thousands of duplicate image files clogging their shared digitisation databases. The problem, years in the making, has been thrust onto the agenda by a July 1 deadline tied to the Grand Paris Numérique interoperability framework, which requires participating institutions to submit clean, deduplicated asset libraries before the end of the third quarter.

The timing is not accidental. Since the Paris 2024 Olympics concluded, city and regional authorities have been funnelling fresh resources into legacy activation projects, including expanded public access to digitised archives as part of the Seine-Saint-Denis cultural uplift programme. Duplicate records directly slow that effort: when the same photograph or engraving appears under three different catalogue numbers, curators cannot confidently link items to exhibitions, researchers waste hours chasing phantom references, and the public-facing portals display confusing or conflicting metadata.

Where the Backlog Is Worst

Two institutions are bearing the brunt of this week's remediation push. The Bibliothèque historique de la Ville de Paris, on Rue de Sévigné in the Marais, has confirmed it is working through a backlog of image files accumulated during a rapid scan drive conducted between 2021 and 2023. Staff are using open-source perceptual hashing tools to flag near-identical files before a manual review layer signs off on deletions. Separately, the Musée Carnavalet, also on Rue de Sévigné and the city's primary museum of Parisian history, is cross-referencing its digitised photograph collection against the shared Joconde national catalogue managed by the Ministry of Culture to eliminate entries that were uploaded twice during system migrations.

The Grand Paris Express construction project has added an unexpected dimension. As archaeological finds along the new metro lines — particularly around the Saint-Denis Pleyel hub — were photographed on-site and then re-photographed in laboratory conditions, duplicate image pairs entered multiple institutional databases simultaneously. Coordinators at Inrap, the national preventive archaeology body, are now working with Paris Musées, the umbrella body overseeing fourteen municipal museums, to reconcile those records under a single canonical file per object.

Cost and Scale of the Problem

The practical stakes are considerable. Cloud storage costs for the Paris Musées network rose to approximately €340,000 in 2025, according to budget documents presented to the Paris City Council cultural affairs committee in March 2026. Administrators privately acknowledge that a meaningful portion of that expenditure covers redundant data. Industry benchmarks for large institutional digitisation projects suggest duplicate rates of between eight and fifteen percent are common without active deduplication protocols — applied to a collection of the scale held across Paris's fourteen municipal museums, that translates to a potentially significant drag on both storage bills and catalogue integrity.

The problem is not unique to Paris. The British Library and the Rijksmuseum in Amsterdam have both published case studies on deduplication workflows in recent years, and their methodologies are informing what Paris Musées is now adapting locally. What makes the Paris situation more acute is the density of institutions sharing overlapping collections within a relatively small geographic footprint — the Marais alone houses several major archives within walking distance of each other — and the relatively short runway before the Grand Paris Numérique compliance deadline.

For researchers and members of the public using the Paris Musées Collections portal, the practical advice this week is straightforward: if a search returns what appear to be identical results under different catalogue numbers, use the feedback button on each record to flag the discrepancy. The Paris Musées digital team has confirmed it is actively monitoring those reports as part of the current clean-up sprint. Institutions expect a first round of bulk deletions and file merges to be completed by September 30, with a public update on progress scheduled for the autumn heritage season. The target is a cleaner, faster, and genuinely searchable archive — one that can support the expanded access programmes tied to the ongoing Seine urban regeneration agenda without being undermined by its own filing system.