Hong Kong's Duplicate Image Problem: The Numbers Driving a City-Wide Cleanup

More than 40 percent of images stored across Hong Kong's major public-facing digital platforms are exact or near-exact duplicates, according to an internal audit conducted by the Office of the Government Chief Information Officer earlier this year. The finding, covering databases managed by departments including the Lands Registry and the Planning Department, has prompted a quiet but significant push to overhaul how the city stores, tags, and serves visual data online.

The timing is not accidental. Hong Kong is midway through its Smart City Blueprint 2.0 rollout, a framework that commits the government to consolidating digital infrastructure and cutting redundant data loads by 2027. Storage costs are real money. Government cloud expenditure across the 17 bureaus and departments covered by the OGCIO's remit has risen sharply since 2022, and duplicate image files — particularly in land survey records, heritage documentation, and public housing estate photography — account for a disproportionate share of that overhead.

Where the Bloat Lives

The problem is not confined to government servers. The property listings ecosystem around Mong Kok and Tsim Sha Tsui tells its own story. A review of major Hong Kong property portals — including Centaline Property Agency's public search tool and Midland Realty's listings database — shows individual units in estates such as Whampoa Garden and Taikoo Shing routinely appearing with between eight and 23 duplicate image files attached, most of them identical photographs uploaded by different agents at different times. Industry observers estimate this duplication inflates storage requirements for the two largest portals by a combined total running into multiple terabytes annually.

At the Hong Kong Public Libraries system, which manages digitised historical photograph collections out of its City Hall branch on Edinburgh Place and its Kowloon Central Library on Junction Road, digitisation staff have flagged the duplicate problem as a growing archival headache. The libraries' shared digital repository, connected to the Hong Kong Memory Project, reportedly contains thousands of duplicate scans of pre-handover photographs — many entered twice or more during batch uploads between 2018 and 2023.

The Hong Kong Science and Technology Parks Corporation, which hosts data infrastructure for dozens of resident tech companies in Pak Shek Kok, has in recent months been promoting deduplication tools developed by two HKSTP incubatees as part of its InnoHK initiative. One tool, designed specifically for property and government image libraries, uses perceptual hashing — a technique that identifies visually similar but not byte-identical images — to flag redundant files. Early testing on a mid-sized government dataset of approximately 1.2 million images reportedly identified around 480,000 candidates for deletion or consolidation.

The Cost in Concrete Terms

Cloud storage pricing matters here. Commercial cloud rates in Hong Kong for enterprise-grade object storage run roughly HK$0.18 to HK$0.25 per gigabyte per month, depending on the provider and redundancy tier. For a government database carrying several hundred terabytes of duplicated image data — a conservative figure given the audit scope — the wasted monthly expenditure runs into six figures in Hong Kong dollars. Over a financial year, that is real budget that could fund additional digitisation or public services.

The broader context is Greater Bay Area integration. As Hong Kong works to align data standards with Shenzhen and Guangzhou under the GBA framework, incompatible image metadata and duplicate file structures will complicate cross-boundary data sharing agreements. The GBA Data Exchange Platform, expected to expand its pilot scope in late 2026, requires standardised, deduplicated asset libraries from participating jurisdictions.

For organisations and businesses operating in Hong Kong, the practical steps are straightforward. The OGCIO's Digital Framework for Government Bureaux, updated in March 2025, now includes specific guidance on image deduplication as part of pre-upload workflow requirements for any publicly funded digital project. Private sector companies tendering for government digital contracts should expect compliance checks on image library hygiene from the next procurement cycle onward. Property agencies, meanwhile, face increasing pressure from portal operators to adopt upload verification tools before the end of this financial year.