Delhi's Digital Archives Race to Fix a Duplicate Image Crisis — Here's What Happened This Week

The Delhi Urban Shelter Improvement Board confirmed this week it had flagged more than 4,000 duplicate image files inside its beneficiary records database, a problem that has quietly compounded since the agency digitised housing-scheme applications during the 2022–23 fiscal push. The discovery, surfaced during a routine audit ahead of a July 15 system migration deadline, is forcing staff across three DUSIB zonal offices — in Rohini, Dwarka, and Saket — to manually verify thousands of beneficiary photo IDs before the new database goes live.

The timing matters. Delhi is in the middle of Phase 4 expansion work on the Metro, with land-acquisition paperwork for the Janakpuri West–RK Ashram corridor already under administrative strain. Any systemic failure in government image databases ripples quickly into documents that residents need to access housing benefits, relocation compensation, and identity verification tied to resettlement schemes. A duplicated or wrongly tagged photograph in one file can delay a family's entitlement for months.

Where the Backlog Is Hitting Hardest

The problem is not confined to DUSIB. The Delhi State Archives, located on Shyam Nath Marg near Civil Lines, has been running a separate digitisation drive covering pre-Partition municipal records and heritage photographs of Old Delhi neighbourhoods including Chandni Chowk and Ballimaran. Archivists there encountered a parallel issue: an automated batch-upload tool used in January 2026 created mirror copies of roughly 1,200 image files, some tied to irreplaceable glass-plate negatives. Staff have been working since May to reconcile the duplicates using hash-verification software, but as of this week the reconciliation was still incomplete.

At the Indraprastha Institute of Information Technology Delhi — IIIT-Delhi, based in Okhla Phase III — researchers who consult with city agencies on records management say the root cause is consistent across departments: bulk-scanning contracts awarded to private vendors often lack standardised file-naming protocols. When different vendors hand over deliverables using different folder structures, the receiving agency's system generates duplicate entries the moment files are merged. IIIT-Delhi's data systems group flagged this structural gap in a January 2026 workshop report shared with the Delhi government's IT department, recommending mandatory SHA-256 checksums on all image deliverables — a step that had not been uniformly adopted by the time this week's audit surfaced DUSIB's backlog.

What the Data Shows

Figures from the Ministry of Electronics and Information Technology's 2025 e-Governance survey — covering state-level digital records programmes — found that image-related data errors, including duplicates and mislabelled files, accounted for roughly 18 percent of grievances filed against digitised beneficiary schemes nationally. Delhi ranked among the top five states by volume of digitised records, with over 3.2 crore document images uploaded to various departmental portals between 2021 and 2025. That scale makes even a small duplication rate statistically significant: an 18 percent error rate across millions of files translates into hundreds of thousands of potentially compromised records.

The AAP administration announced in its 2025–26 budget a ₹47 crore allocation for e-governance infrastructure upgrades, a portion of which was earmarked for database quality assurance. Whether the DUSIB and Archives issues fall within the scope of that spending has not been publicly clarified by the IT department.

For residents whose documents are caught in the backlog, the practical advice from legal aid workers at the Delhi Legal Services Authority on Patiala House Courts Road is to retain all original physical documents — Aadhaar cards, ration cards, tenancy agreements — and to request a written acknowledgment from the relevant zonal office if a digital application appears stuck. DUSIB's helpline has reportedly been directing callers to zonal offices rather than resolving queries centrally, meaning Rohini and Dwarka offices in particular are seeing longer queues for document-status checks.

The July 15 migration deadline is not expected to shift. That gives database teams roughly ten working days to clear the image backlog — a tight window that archivists and IT staff across several agencies described this week as uncomfortable but not impossible, provided vendor cooperation holds.