Melbourne's Duplicate Image Problem: How the City Stacks Up Against London, Amsterdam and Singapore

Melbourne's public image collections are riddled with duplicates. The City of Melbourne alone holds tens of thousands of digitised photographs across its archives on Little Collins Street, and cultural sector workers say a significant portion of those records flag the same image under multiple catalogue entries — wasting storage, confusing researchers and eroding the reliability of public collections.

The issue has landed back on the agenda this week after Victoria's Public Record Office issued updated guidelines for state agencies, pushing institutions to adopt automated deduplication before the end of the 2026–27 financial year. The directive applies to councils, state-funded galleries and libraries, pulling in bodies from the State Library of Victoria on Swanston Street to regional collections in Ballarat and Geelong.

What Other Cities Have Already Done

Amsterdam's Rijksmuseum completed a full deduplication sweep of its online collection in 2023, using perceptual hashing tools developed with the Dutch Digital Heritage Network. The project cleared roughly 12,000 redundant entries from a catalogue of 700,000 items. Singapore's National Heritage Board took a different route, contracting a local AI firm in 2024 to cross-reference holdings across six institutions simultaneously, cutting manual review time by an estimated 60 percent, according to figures the Board published in its 2024–25 annual report.

London's Victoria and Albert Museum began the same process in 2022, prioritising its South Asian textile collection — a category where donated items had often been photographed multiple times under different curatorial regimes. By 2025 the V&A had resolved more than 8,000 flagged duplicates across that single collection category alone.

Melbourne is behind. The State Library of Victoria, which holds more than two million images in its Pictures Collection, is still working through a pilot deduplication project that began in late 2024. A broader rollout has not yet been publicly scheduled.

Why Melbourne's Situation Is Complicated

Unlike Singapore or Amsterdam, Melbourne's institutional holdings are split across dozens of bodies with no single coordinating authority for digital standards. The Australian Institute of Aboriginal and Torres Strait Islander Studies maintains its own protocols for First Nations image collections, which impose separate consent and access rules that complicate automated scanning. The Immigration Museum on Flinders Street holds community-donated photograph collections from more than 30 diaspora groups, many catalogued by volunteer archivists using inconsistent metadata schemas.

That fragmentation has real costs. Cloud storage for large image libraries runs at roughly $0.023 per gigabyte per month on standard government procurement rates through the Victorian Government's whole-of-government agreement with major providers — a figure that compounds when thousands of duplicate high-resolution files sit undetected across multiple institutions for years.

The City of Melbourne's digital strategy unit, based in Council House 2 on Little Collins Street, has been trialling open-source deduplication software since March 2026 across a subset of its urban planning photography archive. Council documents tabled in May 2026 describe the pilot as covering approximately 40,000 images from the Hoddle Grid redevelopment series captured between 2018 and 2023.

The Victorian government's digital modernisation agenda, flagged in the 2025–26 state budget under the Digital Strategy for Government program, allocated $4.2 million to public records infrastructure, though how much of that flows to image deduplication specifically has not been publicly itemised.

For researchers and journalists, the practical consequence of uncleaned archives is straightforward: searches return false positives, licensing teams process the same rights clearance twice, and historians working on Melbourne's built environment or migration history spend hours confirming whether two catalogue records represent the same photograph or a near-identical twin.

Public Record Office Victoria's updated guidelines require agencies to submit a deduplication compliance report by 30 June 2027. Institutions that cannot demonstrate progress by that date face a mandatory audit. For smaller councils and regional galleries operating on tight digital budgets, the deadline is already a pressure point — and Melbourne's cultural sector will be watching whether the State Library's pilot, now entering its second phase, produces a replicable model fast enough to matter.