Software System to check compliance of Packaged Commodities under Legal Metrology(Packaged Commodities) Rules, 2011 by scanning products, images and labels.
Ministry of Consumer Affairs, Food & Public Distribution · Agriculture, FoodTech & Rural Development · Software
Statutory rules define correctness for you, training data is a supermarket away and any judge can test it with the packet in their hand — just do not dodge the font-size-in-millimetres problem, because that is the only part of this that is actually hard and therefore the only part that distinguishes you.
Data: Legal Metrology Act and Rules reference: https://consumeraffairs.gov.in/pages/legal-metrology-act — together with the Legal Metrology (Packaged Commodities) Rules, 2011
What it actually is
Every packaged product sold in India must legally print certain things on it — who made it, how much is inside, the maximum price, when it was packed — in a prescribed format and a minimum letter size. Inspectors currently check this by hand, one packet at a time. The ask is software that photographs a label and reports which of those legal requirements it fails.
What to build
A label compliance scanner where an inspector photographs a package and receives a rule-by-rule verdict: text detection and OCR extracts the printed declarations, a matcher identifies which mandatory declarations under the 2011 Rules are present and which are missing, a geometric analysis converts detected character heights into millimetres using package dimensions or a reference object so the prescribed minimum font size can actually be checked rather than estimated, format validation catches malformed MRP and net quantity declarations, and the result is written into a compliance report with the offending regions annotated on the photograph, filed into a searchable product repository with inspection history and an enforcement dashboard.
Smallest thing that wins the room
Let a judge hold up a snack packet from their own bag, photograph it, and get back a report naming the specific declaration that is missing or under-size with that region circled on the image.
How crowded this one gets
A guess, projected from the 2025 statements — the last year where both the submission counts and the winners were published.
Quieter than 52% of the 226 · #109 of 226 by expected field
A normal-sized field. Your idea has to be good, not miraculous.
Why: central ministry statements sat below the average.
This is a guess, not a fact
Nobody has published 2026’s numbers yet. This is an analysed estimate from last year’s pattern, so please do not take it as the truth — check the live counter on the SIH portal before you decide anything. The range covers the middle half of likely outcomes, so one statement in two lands outside it. Entry closes at 500 ideas per statement, so no range goes past that — a statement that reaches the cap fills and shuts rather than drawing an unlimited crowd. The model reads only three things a team can see before choosing — software or hardware, the theme, and what kind of body posted it — and those explain about a quarter of the variation in last year’s field sizes (R² 0.25 on held-out statements). Trust the band more than the number, and the ordering more than either. It cannot see how good your idea is, which is the part that actually decides it.
The scores
The number is the shorthand. The line under it is the reason.
Acceptance potential
4/5An underrated gem — unlimited free training data, a statutory definition of correctness that removes all requirement ambiguity, a demo that works on anything a judge produces, and a legal metrology enforcement domain that essentially no team will gravitate toward.
Feasibility
4/5Training material is limitless and free — every packaged product in any shop is a sample — and the compliance criteria are not a matter of judgement but a published statutory rule set you can encode exactly, which is a far better position than the usual case of inventing your own definition of correct.
Innovation scope
3/5What must be checked is fixed by the Rules and the functional requirements are enumerated, but how you solve the genuinely hard part — recovering real-world millimetre font height from a photograph of a curved or crumpled package — is completely undetermined and is where the actual engineering lives.
Clarity
5/5Fully specified: the mandatory declarations are named, the checks are enumerated including font size and placement, the report formats, repository, role-based access and dashboard are all listed, and the statement links the governing rules so the definition of compliance is external and unambiguous.
Effort
HeavyText detection and OCR on curved reflective packaging, a rule engine encoding the 2011 Rules, a metric font-size estimation pipeline, report generation, a searchable repository and an enforcement dashboard is five components, and the font-size geometry alone will absorb more time than the OCR.
Demo-ability
EasyA judge can hand you an arbitrary product from their own pocket and check your verdict against the packet in their hand, which makes this one of the very few statements where the demo is both instant and independently verifiable.
In its favour
- Green flag: Correctness is defined by a published statutory rule set rather than by your judgement, so you cannot be accused of choosing a convenient definition of compliance and the Rules are linked in the statement itself
- Green flag: Training data is effectively unlimited and free — one trip to a supermarket produces hundreds of labelled samples across every packaging type you will encounter
- Green flag: The demo works on any product a judge produces, which is a far stronger position than a curated test set you prepared
- Green flag: Legal metrology enforcement is an obscure domain that will attract almost no competing teams despite being a genuinely well-posed problem
Against it
- Red flag: Font size is the requirement everyone will skip because it is the only genuinely hard part — converting character height in pixels to millimetres on a curved, glossy, foil packet needs real geometry, and a team that reports every other check but fingers-crosses this one has dodged the interesting problem
- Red flag: The Rules prescribe font size relative to the principal display panel area, so you need the package's real dimensions, which means either a reference object in frame or a declared pack size — decide which and make it part of the workflow
- Red flag: Reflective foil, curved bottles and multilingual labels break scene text detection in ways that flat cardboard test images will never reveal
- Red flag: Declaring a real product non-compliant has legal consequences for a manufacturer, so false positives matter far more than false negatives here and your system should be built to abstain rather than guess
What you will be writing
- DBNet / CRAFT scene text detection on packaging
- PaddleOCR multilingual label extraction
- homography-based pixel-to-millimetre font height estimation
- rule engine encoding Packaged Commodities Rules 2011
- annotated PDF violation report generation
- PostgreSQL product and inspection repository
- Legal metrology enforcement
- Document and scene text OCR
- Consumer protection
Prior art to read before you start
packaged label declaration compliance checking · scene text extraction from product packaging · rule-based regulatory validation
Analysed by Claude Opus. Every score above is a judgment call with its reasoning attached — kindly cross-check this against the official statement on the SIH portal before your team commits to it.