Paris Takes a Hard Look at Its Duplicate Image Problem — and How It Stacks Up Against London and Berlin

The Bibliothèque nationale de France holds more than 20 million photographic items. The Paris Musées network — covering 14 municipal museums from the Musée Carnavalet to the Petit Palais — went fully open-access in 2020, releasing roughly 150,000 digitised works into the public domain. Now, six years on, curators and archivists across the capital are confronting a messy downstream consequence: the same image, often of the same Haussmann facade or Seine quayside, appearing in dozens of separate institutional records, tagged differently, sized differently, and sometimes attributed to different photographers entirely.

This is the duplicate image problem. It sounds bureaucratic. It has real costs.

When public collections overlap without coordination, search tools degrade, rights clearance becomes a labyrinth, and the cultural record — the kind that feeds journalism, academic research, and the city's own tourism marketing — quietly rots with inaccuracy. The issue has sharpened in 2026 because the Grand Paris Express, the largest infrastructure project in Europe by some measures, is generating tens of thousands of construction and heritage documentation photographs per month across 68 planned stations. Multiple agencies — Île-de-France Mobilités, the Société du Grand Paris, and individual arrondissement heritage offices — are each logging images of the same sites with no shared deduplication protocol in place.

What Paris Is Doing — and Where the Gaps Are

The BnF's Gallica platform introduced an automated similarity-detection layer in late 2024, using perceptual hashing to flag near-duplicate scans within its own holdings. It was a meaningful internal step. The problem is jurisdictional: Gallica's system does not talk to Paris Musées' database, which in turn does not communicate with the Institut national de l'audiovisuel or the documentary holdings of the Hôtel de Ville. Each institution runs on different metadata standards — some use Dublin Core, others LIDO — and there is no citywide data-sharing agreement compelling them to reconcile records.

London moved earlier. The British Library, working alongside the Victoria and Albert Museum and the Museum of London, piloted a cross-institutional deduplication framework in 2023 under the UK's DCMS-funded Culture and Heritage Capital programme. By early 2025 the pilot had flagged more than 80,000 duplicate or near-duplicate image entries across participating collections, according to a progress report published by the programme. The V&A alone removed or merged approximately 12,000 redundant catalogue records as a result.

Berlin has taken a different route. The Stiftung Preußischer Kulturbesitz — the foundation that oversees the Prussian Cultural Heritage collections including the Pergamon and the Staatsbibliothek — adopted a centralised image management system in 2022 that mandates deduplication before any new digital asset enters the shared repository. It is a stricter model, closer to a publishing pipeline than a retrospective clean-up, and officials there have been candid that retrofitting older collections remains unfinished work.

The Stakes in a City Rebuilding Its Own Image

Paris has particular reason to get this right in 2026. The post-Olympics urban legacy programme, centred on the Seine-Saint-Denis corridor and the revitalised berges along the Quai d'Austerlitz, is generating a second wave of institutional documentation. The Agence nationale pour la rénovation urbaine is co-funding heritage photographic surveys in Aubervilliers and Saint-Denis as part of rehousing impact assessments. If those surveys feed into archives already cluttered with duplicates, the evidentiary chain for future heritage and planning decisions becomes unreliable.

Rental market pressure in inner arrondissements — median rents in the 10th and 11th crossed €30 per square metre for small furnished apartments in the spring of 2026, according to data from the Observatoire des Loyers de l'Agglomération Parisienne — has also pushed property documentation photography into overdrive, with landlords, notaires, and platform operators each building their own image stores of the same streets on the Canal Saint-Martin and around the Marché d'Aligre.

The practical path forward likely involves the Direction des Affaires Culturelles de la Ville de Paris brokering a minimum metadata agreement between the city's own agencies before attempting any external harmonisation. Several European peers, including Amsterdam's Stadsarchief, have managed this at municipal level without waiting for national policy. Paris has the institutions, the technical capacity, and — after 2024 — a demonstrated willingness to open its collections to the world. The missing piece is coordination, and that is a political choice, not a technical one.