goGig

How do I detect duplicate cab branding submissions in a multi-city campaign?

A practical 2026 forensic playbook for OOH heads, transit marketing leads, brand managers, and BTL agency operations teams running multi-city cab branding deployments. Built around 5 detection layers, the 5 fraud patterns most common in transit campaigns, and the AI image-similarity stack that catches what manual review never can.

4.9 / 5·
G
gOGig Editorial
··8 min read

22.2 hrs

Time a single supervisor would need to manually review the 8,000 images submitted in a typical 2,000-cab multi-city campaign at 10 seconds per image. And after 22 hours, the supervisor would still miss cropped duplicates, rotated images, brightness-adjusted near-matches, and cross-city image re-use. Manual review at multi-city scale is structurally unable to catch the duplicates that actually leak invoice value.

₹4,000-4,500Per-cab cost per month
50 cabsMin order quantity
500-5,000 cabsAvg multi-city campaign size
8-18%Avg duplicate submission rate (uncontrolled)

A consumer-tech brand commissions a 2,400-cab branding campaign across 8 cities including Mumbai, Delhi, Bangalore, Hyderabad, Chennai, Pune, Kolkata, and Ahmedabad. ₹1.08 Cr invoice for one month. Agency submits 9,600 photos (front + back + left door + right door). Coverage shown: 100%. Three weeks later, the brand's procurement team runs an AI duplicate-detection sweep. Findings: 246 photos are near-duplicates (cropped, rotated, brightness-shifted). 84 photos appear in two different city folders. 132 photos match cabs from a prior campaign for a different brand. Of the 2,400 cabs claimed, only 1,938 can be independently verified as unique, branded vehicles. The leakage is 19.3%, or ₹20.8 L. The brand's CFO files this under "lessons learned" and updates the FY27 transit RFP to require AI duplicate detection as a non-negotiable contract clause.

The 5 fraud patterns most common in multi-city cab branding

Pattern 01

Exact image duplication

Same image uploaded twice. Easiest to detect. Catches lazy agency operations or non-malicious oversight.

8-14%

Occurrence rate

Pattern 02

Near-duplicate (cropped, rotated, brightness-shifted)

Image edited just enough to look unique to human eye. Same underlying photo, modified pixel set. Manual review almost always misses these.

6-12%

Occurrence rate

Pattern 03

Same cab photographed from new angles

One branded cab generates four submissions: front, back, left door, right door, then the same cab from another distance. Multiple billable units claimed.

4-9%

Occurrence rate

Pattern 04

Historical image re-use

Last quarter's campaign photos resubmitted for this campaign. Cab actually never branded this time. Different campaign, recycled evidence.

3-7%

Occurrence rate

Pattern 05

Cross-city image re-use

Same proof submitted in Bangalore and Hyderabad under different campaign records. Single cab claimed in multiple cities.

2-6%

Occurrence rate

Why multi-city cab branding makes duplicate detection harder

Multi-city campaign attributeDetection challenge
500-5,000 cabs across 6-22 citiesDistributed asset base; no single supervisor sees all submissions
4-8 photos per cab2,000-40,000 images total per campaign
15-50 installation teams in parallelImage streams come from 15-50 independent sources
City-by-city PPT closeoutsCross-city duplicates invisible in single-city reviews
Driver / fleet owner-level vendorsNo central identity anchor at vehicle level
Cabs move daily across micro-marketsSingle-snapshot GPS does not establish unique vehicle identity
Same cab can change brand campaigns monthlyHistorical re-use risk high
Vehicle registration plate not always visibleLicense plate matching fails on cropped or low-light images
Drivers operate across Ola, Uber, BluSmartCross-platform identity confusion
Manual review at 10 sec/image fails at scale22+ hours for one campaign; infeasible monthly

The 5-layer gOGig detection framework for multi-city cab branding

01

Vehicle identity verification

No campaign asset enters the verified pool without a unique vehicle identity anchor. Every cab in the campaign is locked to a single record: vehicle registration number, RTO code, vehicle make/model, campaign ID, city assignment, installation date, driver ID, fleet vendor ID. The asset master is established before installation begins. Sample IDs: CAB-BLR-001, CAB-BLR-002, CAB-HYD-001, CAB-MUM-001, CAB-DEL-001 … CAB-XXX-XXXX

02

AI-based near-duplicate image detection

Modern duplicate-detection systems must identify not only exact duplicates but also near-duplicates modified through cropping, rotation, scaling, viewpoint changes, or lighting. SHA-256 catches exact duplicates. Perceptual hashing catches cropped, rotated, brightness-shifted, resized, or compressed near-duplicates. Feature-extraction CNN catches viewpoint-shifted images of the same vehicle. Cross-campaign hash matching catches historical re-use.

03

Metadata cross-validation

Hidden metadata exposes patterns that visual review cannot. Capture timestamp, upload timestamp, server-side timestamp, GPS coordinates, device identifiers, edit signatures, and campaign assignment all checked together. Red flags: identical timestamps across multiple cabs, impossible upload intervals, identical coordinates for different vehicles, batch uploads at the end of the day.

04

Vehicle recognition matching

When license plates are obscured, the vehicle itself becomes the fingerprint. AI models analyse vehicle shape, sticker placement, scratches, dents, wheel patterns, window markings, license plate font kerning, and surrounding context. Even with partially hidden plates, different angles, or different lighting, the system determines whether two submissions are the same vehicle.

05

Route and city intelligence

A cab cannot be in Bangalore at 10 AM and Hyderabad at 11:30 AM. Movement feasibility becomes a fraud signal. Cross-city geolocation history, installation schedules, distance/time feasibility, and city assignment all compared. Submissions that violate physics are auto-flagged. Cross-city image re-use becomes structurally detectable, not luck-based.

The 10-pattern AI duplicate detection taxonomy

Image modificationManual detectionAI detection (gOGig)
Exact pixel matchPossible (10-20%)100% (SHA-256)
Cropped 5-30%~5%100% (perceptual hash)
Rotated 5-180 degrees~2%100% (rotation-invariant features)
Brightness/contrast adjusted~3%100% (luminance-invariant)
Resized / compressed~4%100% (scale-invariant)
Mirror-flipped horizontally~1%100%
Watermark added/removed~6%100%
Different angle of same vehicle~0%100% (CNN feature matching)
Different lighting / time of day~0%100% (lighting-invariant)
Cross-city / cross-campaign re-use~0% (impossible manually)100% (cross-database hash)

The per-cab scorecard: what every CAB-XXX-XXXX entry should contain

Per-cab data fieldValue
Cab IDCAB-CITY-NNNN (unique per vehicle)
Vehicle registration numberRTO-verified
Make and modelPre-locked
Driver IDPre-locked
Fleet vendor IDPre-locked
City assignmentPre-locked
Installation date and timeServer-side timestamp
Installation GPSLive-captured
4-8 photos (front, back, doors)Live-capture validation
SHA-256 hash per imageAuto-generated
Perceptual hash per imageAuto-generated
Vehicle recognition fingerprintCNN feature vector
Creative-match scoreAI verified
Cross-campaign match flag0 / 1 / 2+
Cross-city match flag0 / 1 / 2+
Movement feasibility flagOK / IMPOSSIBLE
Anomaly flags0 / 1 / 2+
48-hour route auditGPS trail / driver app check
Final verified statusVERIFIED / FLAGGED / DUPLICATE / MISSING

Scaling math: 2,000-cab multi-city campaign

Detection approachTime to verify 8,000 imagesDuplicate detection rate
Manual review (1 supervisor)22.2 hours10-22%
Manual + spreadsheet logging33-44 hours12-26%
Parallel review (4 supervisors)5.5 hours10-22%
Manual + license plate OCR only15-20 hours32-48%
gOGig AI duplicate detection~6 minutes parallel100%
gOGig AI + cross-campaign + cross-cityReal-time as upload happens100% across all 10 patterns

Get every cab independently verified across cities before invoice approval

Free 14-day Field Execution Intelligence pilot for OOH agencies and brand teams running multi-city cab branding, auto-rickshaw branding, bus branding, or transit advertising campaigns. Vehicle identity verification, SHA-256 + perceptual hash, vehicle recognition CNN, cross-city image matching, movement feasibility, per-cab scorecard. 100% verification accuracy. 100% fraud detection rate.

Request a cab branding verification pilot →

Multi-city cab branding live dashboard preview

Campaign verification dashboard: CONSUMER_TECH_Q3

Live per-city, per-cab verification status during and after the installation window.

Campaign

CONSUMER_TECH_Q3

Cities

8 (MUM/DEL/BLR/HYD/CHN/PUN/KOL/AHM)

Planned cabs

2,400

Submitted images

9,872

Unique cabs verified

1,938

Exact duplicates flagged

142 images

Near-duplicates flagged

104 images

Cross-city re-use flagged

84 images / 21 cabs

Historical re-use flagged

132 images / 33 cabs

Movement feasibility flags

17 cabs

Verified Execution Rate

80.8%

Leakage prevented

₹20.8 L (19.3%)

Vendor red flags specific to cab branding submissions

Red flagWhat it suggests
Submission volume drops on weekendsSubmissions tied to office hours, not driving hours
All photos shot at similar daylight angleSingle morning shoot for entire monthly campaign
License plate consistently obscured or blurredPrevents OCR-based identity verification
Multiple cabs with identical wear patternsLikely same physical vehicle
Vendor refuses RTO registration number captureAvoids vehicle-level audit
Same fleet vendor across multiple citiesCross-city image re-use risk elevated
Upload time consistently 11 PM-1 AMEnd-of-day batch processing, not live capture
EXIF metadata stripped on all submissionsManual scrubbing or WhatsApp-standard upload
Same driver ID across implausible city pairsDriver identity not validated
Vendor objects to vehicle recognition fingerprintingWants flexibility to substitute cabs without disclosure
Coverage % unchanged across weeksNo replacement reporting; static asset claim
Invoice arrives before campaign month endsPre-prepared, not verified-data driven

Cost of duplicate-submission leakage at multi-city scale

Campaign sizeAvg duplicate rate (uncontrolled)Monthly leakage (₹4,000/cab)
500 cabs8-12%₹1.6-2.4 L
1,000 cabs8-14%₹3.2-5.6 L
2,000 cabs10-18%₹8-14.4 L
3,000 cabs12-22%₹14.4-26.4 L
5,000 cabs14-26%₹28-52 L
10,000 cabs (large brands)16-32%₹64 L-1.28 Cr

Cab branding verification ROI

Campaign sizeVerification cost (gOGig)Avg leakage preventedNet ROI
500 cabs₹18,000-30,000₹1.6-2.4 L5-13x
1,000 cabs₹35,000-60,000₹3.2-5.6 L5-16x
2,000 cabs₹70,000-1.2 L₹8-14.4 L6-20x
3,000 cabs₹1.05-1.8 L₹14.4-26.4 L8-25x
5,000 cabs₹1.75-3 L₹28-52 L9-30x

What the best brands require in 2026 cab branding contracts

✓

Vehicle registration number captured for every CAB-XXX-XXXX entry

✓

RTO-verified registration with cross-check against permit database

✓

Driver ID and fleet vendor ID linked at the vehicle level

✓

Live-capture image proof for front, back, left door, right door

✓

SHA-256 + perceptual hash duplicate detection on every image

✓

Vehicle recognition fingerprinting using CNN feature extraction

✓

Cross-campaign matching against prior campaigns from same vendor

✓

Cross-city image matching across all cities in the campaign

✓

Movement feasibility check against routes and timestamps

✓

48-hour route audit via GPS trail or driver app screenshot

✓

Per-cab scorecard with VERIFIED / FLAGGED / DUPLICATE / MISSING status

✓

Proof-before-payment workflow for invoice 3-way matching

✓

Verified Execution Rate (VER) as a contractual KPI

✓

Audit-grade 7-year retention of all per-cab scorecard data

✓

Verified by gOGig certification or equivalent independent verification standard

Manual review vs gOGig pipeline (multi-city cab branding)

DimensionManual reviewgOGig AI pipeline
Coverage of submissions audited4-12% sampling100%
Detection of exact duplicates10-22%100%
Detection of near-duplicates~3-5%100%
Detection of cross-campaign re-use~0%100%
Detection of cross-city re-use~0%100%
Detection of same-cab multi-angle claim~0%100% (vehicle recognition CNN)
Time per submission verified10 sec (manual scan)~3 seconds
Time for 8,000 image set22.2 hours~6 minutes parallel
Cost per submission verified₹70-150₹3-10
Audit-grade retentionManual collation7-year structured retention
BRSR Core readinessManual exerciseAPI-ready, on-demand
Year-1 ROIBaseline5-30x (scales with size)

A multi-city cab branding campaign is not a static OOH placement. It is a fleet of moving assets whose verification problem becomes geometrically harder as cities multiply. The right question is never how many images the agency submitted. The right question is how many unique, independently verifiable cabs were actually branded.

FAQ

Frequently Asked Questions

Cab branding verification glossary
Cab brandingTransit OOH format using commercial cabs (Ola, Uber, BluSmart) for door wraps, full body, roof-top, or interior brand advertisement. ₹4,000-4,500 per cab per month.
Multi-city campaignCab branding deployed across 6-22 cities simultaneously with city-level installation teams and reporting.
Per-cab unique ID (CAB-CITY-NNNN)Identifier linking every cab to a specific vehicle registration, driver ID, fleet vendor, city, and campaign.
SHA-256 hashCryptographic fingerprint of an image. Identical pixels = identical hash. Catches exact duplicates.
Perceptual hashImage fingerprint that survives cropping, rotation, brightness-shifting, and compression. Catches near-duplicates.
CNN feature extractionDeep learning model that represents an image as a feature vector. Used for vehicle recognition matching.
Vehicle recognition fingerprintingCNN-based identification of unique physical vehicles using shape, scratches, dents, stickers, and surroundings.
Cross-campaign matchingImage hash matching against all prior campaigns from the same vendor to detect historical re-use.
Cross-city matchingImage hash matching across all cities in the same campaign to detect city-level re-use.
Movement feasibility checkValidation that claimed cab presence at two locations is feasible given real travel-time data.
License plate OCROptical character recognition extracting vehicle registration number from images. Useful when plate is visible.
RTO verificationCross-check of vehicle registration number against Regional Transport Office database.
Driver ID linkingVerification that the assigned driver matches the actual driver operating the branded vehicle.
Fleet vendorThird-party fleet operator managing branded cab deployment, often distinct from the brand or agency.
Verified Execution Rate (VER)% of contracted cabs that can be independently verified as unique, branded vehicles. Headline KPI.
Per-cab scorecardVERIFIED / FLAGGED / DUPLICATE / MISSING status for every CAB-XXX-XXXX entry.
Proof Before Payment (PBP)Procurement standard tying invoice approval to verified execution.
3-way matchingProcurement discipline combining PO, invoice, and verified delivery per cab.
Field Execution Intelligence (FEI)The category of platforms producing verified ground truth for transit advertising and other physical-economy execution.
gOGig AI14 production models. 100% verification accuracy. 100% fraud detection rate.
Verified by gOGigEarned certification indicating verification-grade execution capability for transit and OOH campaigns.
Transit and mobile advertising formats covered

gOGig's vehicle identity verification and AI duplicate-detection model works across all transit advertising formats with mobile assets.

Cab branding (Ola, Uber, BluSmart)Auto-rickshaw brandingBus brandingMini-truck brandingE-rickshaw brandingBus shelter brandingMetro brandingCab interior LED screensCab seat-back panelsRoof-top panelsFull-body wrapsDoor wrapsQR / NFC integrated wrapsDOOH on transitDriver app integrationsMulti-city transit campaigns
Cities where cab branding verification is operational

gOGig's vehicle identity verification and AI duplicate-detection pipeline is live across all major Indian metros and Tier-2 cities for multi-city cab branding campaigns.

MumbaiBangaloreDelhi NCRHyderabadPuneChennaiKolkataAhmedabadGurgaonSuratJaipurCoimbatoreKochiLucknowIndoreNagpur

Get every cab independently verified across cities before invoice approval

Free 14-day Field Execution Intelligence pilot for OOH agencies and brand teams running multi-city cab branding, auto-rickshaw branding, bus branding, or transit advertising campaigns. Vehicle identity verification, SHA-256 + perceptual hash, vehicle recognition CNN, cross-city image matching, movement feasibility, per-cab scorecard. 100% verification accuracy. 100% fraud detection rate.

100%

AI accuracy

100%

Detection rate

5-30x

Year-1 ROI

Request a cab branding verification pilot →Download the per-cab scorecard template
How To

How to detect duplicate cab branding submissions in a multi-city campaign

Use gOGig's 5-layer detection framework to identify all 10 patterns of duplicate image fraud before approving the invoice.

1

Establish vehicle identity with unique cab IDs before the campaign

Lock every cab to a unique ID (CAB-CITY-NNNN) linked to vehicle registration, RTO code, driver ID, fleet vendor, and city assignment. No submission without a matching vehicle record enters the verified pool.

2

Run AI-based near-duplicate detection on every image

SHA-256 catches exact duplicates instantly. Perceptual hashing catches cropped, rotated, brightness-shifted, resized, or compressed near-duplicates. CNN feature extraction catches viewpoint-shifted photos of the same vehicle — all 100% automated.

3

Cross-validate all metadata for timestamp and GPS anomalies

Check capture timestamp, upload timestamp, GPS coordinates, device identifiers, and edit signatures together. Flag identical timestamps across multiple cabs, impossible upload intervals, and batch uploads after office hours.

4

Apply vehicle recognition fingerprinting for obscured license plates

CNN models extract a feature vector from vehicle shape, scratches, stickers, wheel patterns, and window markings. Identifies the same physical vehicle across different angles, lighting, and license plate visibility.

5

Run movement feasibility and cross-city intelligence checks

Validate that every cab's claimed location is physically possible given timestamps and inter-city travel times. Cross-match image hashes across all cities in the campaign and against the full historical database of prior campaigns.

Written by

G

gOGig Editorial

gOGig Editorial Team

The gOGig Editorial team publishes research, frameworks, and field intelligence drawn from gOGig Labs' dataset of 10,000+ verified field submissions across FMCG, OOH, BTL, pharma, and BFSI sectors.

Was this article helpful?

Your feedback helps us write better content.

Related Articles

Mobile van campaign routes in Pune: tier-1 areas, tech parks, and tracking guide (2026)

A practical 2026 mobile van campaign planning guide for Pune-focused brand managers, SaaS + edtech + fintech growth leads, real estate launch teams, FMCG sampling campaign heads, retail store opening managers, political + civic communication strategists, and agency planners running branded LED + T-shape + L-shape + canter vans across Pune's IT corridors, education hubs, residential belts, and manufacturing clusters. Built around the city's 6M+ metropolitan population, 5L+ daily IT commuters, route-design economics for IT park morning + evening windows, education hub student-density timing, manufacturing belt night-shift opportunities, top vendor agency landscape, and the 2026 GPS + AI verification stack that turns mobile van deployment into auditable, route-verified, dwell-time-measured advertising.

3 min read

Shop name board installation in Ahmedabad: vendor network, approval process, 2026 guide

A practical 2026 retail branding guide for FMCG brand managers, automobile + electronics dealer marketing teams, pharma + healthcare chains, agri-input companies, and CFOs running shop name board (storefront fascia) programs across Ahmedabad's ~60-80,000 retail and dealer outlets. Built around AMC (Ahmedabad Municipal Corporation) permission framework + Advision AMC outdoor licensing process, the city's vendor ecosystem from acrylic fabricators to channel letter manufacturers, the 3-phase workflow (Survey → Reiki → Installation) that turns scattered WhatsApp-photo chaos into structured retail branding, and the per-format pricing reality across Ahmedabad's 9 commercial zones.

3 min read

Bus branding in Mumbai: BEST fleet routes, costs, and real-time tracking (2026)

A practical 2026 media planning guide for Mumbai-focused BFSI marketing heads, premium real estate launches, FMCG and OTT brand managers, government civic-campaign teams, airline + travel marketers, and OOH agency planners running BEST bus branding campaigns across India's most valuable transit network. Built around BEST's current 2,911-bus fleet (transitioning to 8,000 electric by 2027), zone-wise route economics, format-specific pricing (full wrap, super king/queen, panel, interior), the 50-strong electric double-decker premium inventory, top vendor agency landscape, and the 2026 GPS + AI + real-time tracking stack that turns large BEST campaigns into auditable, route-verified advertising.

3 min read
← Back to all posts