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

Melbourne's public institutions are sitting on a cataloguing headache that archivists in Amsterdam and Toronto have been wrestling with for years: duplicate images embedded deep inside digitised collections, eating storage budgets, confusing researchers and undermining the credibility of public-facing databases. The City of Melbourne's digital archive program, managed through the Melbourne City Archives unit on Little Collins Street, flagged the problem internally after a 2025 audit found hundreds of repeated photographs across its urban planning records alone.

The timing matters. Victoria's government has pushed an aggressive digitisation agenda since 2023, accelerating the transfer of physical council and cultural records to online repositories as part of the broader Digital Victoria framework. More material going online faster means more duplicates slipping through. Libraries, galleries and local councils in Melbourne are now confronting a quality-control backlog at the exact moment they are expanding public access.

What Other Cities Are Doing

Amsterdam's Stadsarchief, the municipal archive that holds more than three million digitised items, rolled out automated deduplication software across its photo collection in late 2023, cutting verified duplicate entries by roughly 18 percent within 12 months, according to figures published on its own institutional reporting page. The tool cross-references pixel hash values against metadata tags — a process that takes seconds per image but required a six-month calibration period before archivists trusted the outputs.

Toronto Public Library took a different approach. Its 2024 Digital Collections Strategy, publicly available on its website, prioritised human-led review for culturally sensitive material, particularly photographs depicting Indigenous communities, before any automated culling. The library assigned four dedicated digitisation quality officers to the task — a staffing level that drew criticism from some quarters as insufficient given the size of the backlog.

Seoul's National Folk Museum used a hybrid model, deploying AI-assisted flagging followed by specialist curatorial sign-off. Its public documentation from early 2025 described the process as cutting unresolved duplicate records from around 12,000 to under 400 across a single photographic collection within one fiscal year.

Melbourne, by contrast, has no single coordinating body managing this work across institutions. The State Library Victoria on Swanston Street runs its own deduplication process inside Trove-linked collections, while the Australian Centre for the Moving Image on Federation Square handles its own film and digital image catalogues separately. The City of Melbourne Archives operates under different software entirely. There is no shared protocol governing how a flagged duplicate gets reviewed, resolved or removed from public view.

The Local Patchwork

For researchers and journalists trying to pull historical images from Melbourne's publicly accessible collections, the practical effect is wasted time and occasional embarrassment. A photograph described as depicting the 1956 Melbourne Olympic construction on one database turns up again on another platform attributed to a different decade and a different suburb. Neither entry is necessarily wrong — but both cannot be right.

The State Library Victoria's Trove presence, one of the most heavily used public archives in the country with tens of millions of items catalogued, has undergone several metadata cleaning rounds. The National Library of Australia, which operates Trove nationally from Canberra, published its most recent data quality report in March 2025, though specific duplicate-rate figures for individual city collections were not broken out in that document.

Victoria's Public Record Office, headquartered in North Melbourne, confirmed in its 2024–25 annual report that it was trialling new ingest standards designed to catch duplicate files before they enter the permanent collection — a prevention model rather than a retrospective clean-up. That distinction matters: fixing duplicates after the fact costs significantly more in staff hours than stopping them at the door.

For institutions still deciding which path to take, the global evidence is pointing in one direction. Automated tools are faster and cheaper for straightforward photographic archives. Human review remains non-negotiable for collections with legal, cultural or provenance sensitivity. Melbourne's fragmented institutional landscape means no single shortcut will work everywhere — but the window to agree on shared standards is narrowing as the volume of newly digitised material keeps growing. The next State Archives audit of Victorian council collections is scheduled for late 2026, and whatever it finds will land at a moment when the gap between Melbourne's approach and those of comparable global cities will be harder to explain away.