vInspect
AI-driven 100% inline print inspection

Every label. Every defect. At 300 m/min.

vInspect is the print inspection platform built for narrow-web and flexible packaging lines: full-web 100% inspection, AI defect classification, inline color measurement and closed-loop control — all on a modern web UI your shift can actually use.

< ΔE°° 1.3
Color accuracy
up to −50%
Wastage
1060 mm
Web @ 300 dpi
300 m/min
Inspection speed
Production figures from the CCL Label · Dornbirn deployment, documented in the Allied Vision print-inspection case study.
vInspect inspection bridge over a running label press
vInspect · in production
Inspection bridge on a running narrow-web press
vInspect · production at scale

More than 10,000,000 meters of label production inspected — every month.

Across deployed lines today, vInspect processes ten million linear meters of printed web per month at full DPI — every label, continuously, in real time.

100%
Inline coverage
24/7
On running presses
AT
In production today
Field proven · CCL Label, Dornbirn (AT)

Inline inspection on shrink-sleeve labels — printed on transparent film.

Built with CCL Label and Chromasens (now Allied Vision): 2 × allPIXA evo 8k color line-scan cameras over a 1060 mm web at 300 dpi, running at 5 m/s (≈300 m/min) — full inline color measurement and defect detection on transparent film, where traditional color control elements can't be used.

< ΔE°° 1.3
Color accuracy
−50%
Setup time
up to −50%
Wastage
“We used to need about 40 minutes for the same job. Now we need less than half an hour. Once the plates are loaded and we start running, we can already tune the colors directly from the image — no more stopping the press to pull a sample sheet. We register everything live, while it's running.”
Dieter Günther
Production Manager, CCL Label GmbH · on-camera in the deployment film
“The key to the solution lies in the calibration process, which allows us to achieve color measurement accuracy with a color deviation of < ΔE°° 1.3. This means that measurement with the spectral handheld measuring device can be replaced.”
Franz Brandl
Managing Director, Virona GmbH · quoted in the Allied Vision case study
Platform capabilities

One platform — from the camera bridge to the analytics dashboard.

The short list. AI defect detection, inline color and closed-loop control each get their own deep-dive below — these are the platform capabilities that round out a vInspect line.

100% full inspection

Every printed element on every label is analyzed — not sampled. Nothing slips between the cracks of statistical sampling or sheet-only QA.

300 m/min at 300 DPI

CUDA-accelerated pipelines hold full inspection resolution at production speed, so QA never becomes the bottleneck on a fast press.

Pixel-precise defect position

Every defect is localized to a precise web position so downstream converting can automatically eject or skip the bad section — automated waste removal, not manual sorting.

Single or multi-camera bridge

Scaled to your web width. Industrial line-scan and area cameras deliver pixel-precise inspection on narrow-web and flexible packaging lines alike.

PDF-based proofs

Generate inspection masters directly from PDF separations. No physical first article required — visualize and approve the expected print before press time.

Smart job database

A central library of jobs and reference patterns enables one-click reuse, repeat orders, and dramatically faster set-up across shifts and machines.

Advanced quality analytics

Reproducibility scoring, material deformation (e.g. shrink films), length control versus target — data the QA lead can hand straight to the customer.

Role-based access control

Administrators, operators and view-only roles (e.g. cutting-machine staff). Right people, right screens, right permissions across the plant.

Engineered in Austria

Mechanics, electronics and integration designed and built in Austria. Long-life industrial hardware, serviced by the people who built it.

Label web running through the vInspect inspection bridge
INK MISS
COLOR ΔE
REGISTER
AI classifier · what it tags
AI defect intelligence

Operators see decisions — not raw pixels.

Our AI classifier doesn’t just find a defect — it tells you what it is. Print marks vs. real ink misses, intentional splices vs. material flaws, register drift vs. die-cut variation. The shift sees a clean, ranked list and acts on what matters.

Classified

Types auto-tagged: ink-miss, register, splice, color, die-cut, contamination.

Ranked

Critical defects float to the top. Cosmetic noise is suppressed by policy.

Learning

Operator annotations feed back into the model — the system gets sharper per shift.

Closed-loop color control

Inspection that doesn’t just measure — it corrects.

vInspect computes per-channel correction values from inline ΔE measurements and pushes them straight to the press control system. Color drift is dampened in real time, before it becomes a waste roll on the floor.

  • Measure
    Inline ΔE per ink channel, sampled across the full web width — every label, every revolution.
  • Compute
    Per-channel correction (more / less ink) with operator-tunable response curves.
  • Apply
    Adjustments delivered to the press control over the signal service — closed loop, no operator click.
Printing pressink channels CMYK + spotvInspect camerasmulti-camera, inlinevinspect_serverΔE / defect classifierClosed-loop controllercorrection valuesPrintSightOEE · trends · wasterunning webΔ correctionset pointprint
cyan
per-channel ΔE
magenta
per-channel ΔE
yellow
per-channel ΔE
AI multi-camera calibration

One camera or four — the operator sees a single, perfect web.

vInspect uses deep-learning calibration to harmonize exposure, gain and color across every camera in the bridge. No seams, no banding, no stitched-together inspection results — even on the widest webs and trickiest substrates.

Cam-to-cam
color & exposure aligned
Per camera
recalibrated on demand
vInspect camera bridge with two area cameras above the running web
AI calibration · cam-to-cam
Uniform color and exposure across every camera in the bridge
How it runs

From PDF separation to plant-wide OEE — one pipeline.

vInspect slots into your existing line through the signal service, a Meilhaus measurement card and a shared volume for reference assets. No forklift upgrade required.

  1. STEP 01

    PDF-based proof

    Upload PDF separations. vInspect generates the inspection master directly — no physical first-article needed before press time.

  2. STEP 02

    Calibrate

    AI multi-camera calibration aligns every camera in the bridge. Lab color calibration sets the ΔE tolerance and emits a certificate.

  3. STEP 03

    Inspect at speed

    100% inline inspection at up to 300 m/min · 300 DPI. AI classifies every defect and routes critical items to the right operator role.

  4. STEP 04

    Close the loop

    Per-channel ink corrections are computed and pushed to the press control system. Drift is dampened in real time, not at the next reject.

  5. STEP 05

    Hand off to automation

    Pixel-precise defect coordinates are forwarded to downstream cutting and converting machines so bad sections are ejected automatically — no manual sorting station, no good labels lost.

  6. STEP 06

    Analyze

    PrintSight rolls every job into plant-level OEE, waste and setup analytics — with ranked optimization recommendations.

Modern operator UI

A web app every operator on every shift can actually use.

Browser-based, multi-client, multi-tab. Operators monitor running jobs, annotate defects and pull historical rolls without installing a thing. Cutting-machine staff get a focused view that only shows what they need to act on.

Real-time

Camera feeds and KPIs streamed across the plant network.

RBAC

Admins, printers and view-only roles. Right access for every job.

Multi-client

Use it from the press cockpit, the QA office, the cutting line — at the same time.

vInspect operator cockpit running on a multi-monitor press station
Operator cockpit · in production
Job · register · ΔE · annotation
multi-client
ROLEAdministrator

Full system configuration · users · jobs · models

ROLEOperator / Printer

Running job, defect handling, calibration runs

ROLEView-only

Cutting-line view · defect dispatch · read-only

PrintSight · optional analytics layer

Every inspected meter becomes plant-wide insight.

vInspect is the main act — PrintSight is the side show that turns what vInspect already knows into KPIs your plant manager and shift leads can act on. OEE per shift, planned-vs-actual on production time, setup time and waste, with 30-day rolling trends per line.

Reporting windows follow the production day (06:00–06:00) and ISO weeks — the same accounting your shift leads already use.
Team-OEE

Mean of Production, Setup and Waste OEE — one headline number per shift.

Production OEE

ΣPlanned vs. ΣActual production time — where lines slip and why.

Setup OEE

ΣPlanned vs. ΣActual setup time — minutes lost between every roll, quantified.

Waste OEE

ΣPlanned vs. ΣActual waste meters — the most expensive ratio in the plant, exposed.

printsight · TeamPulse · last 30 d
example data
Team-OEE & part-OEEs · 30-day window
88.6%
Team-OEE
92.1%
Production OEE
81.4%
Setup OEE
94.3%
Waste OEE
Soll vs. Ist · last 30 d
ΣProduktionszeitSoll 14,820 min·Ist 13,651 min92%
ΣRüstzeitSoll 2,160 min·Ist 2,654 min81%
ΣAbfallmeterSoll 58,000 m·Ist 61,500 m94%
ΣLaufmeter Gesamt → NettoSoll 9.84 M m·Ist 9.56 M m97%
Team-OEE & part-OEEs · 30-day rolling
100908070
Team-OEEProduction-OEESetup-OEEWaste-OEE
Example dashboard · numbers shown are illustrative, not from a customer line.
Watch it run

See vInspect in operation — and hear it from the customer.

Two short clips from the CCL Label · Dornbirn deployment. The production manager on what changed for his shift, and the engineering lead on how the system works. Both with English captions.

Customer voice·0:49
“40 minutes used to mean 40 minutes.”
Dieter Günther · Production Manager, CCL Label GmbH
How it works·1:18
Inline color at 300 DPI · 5 m/s · 1 m web
Franz Brandl · Virona GmbH

Recorded on the CCL Label · Dornbirn line. German audio, English captions. Used with permission.

Who builds vInspect

A two-person partnership — print expertise meets machine learning.

Long-standing print-industry consulting on one side. Years of building high-performance machine learning and computer-vision systems on the other. Both shipping the same product.

Franz Brandl
Print industry · operations · hardware
Franz Brandl
Virona GmbH
Munich, Germany

Senior print-industry expert and consultant. Long-standing experience advising label and packaging converters on print quality, machine integration and process optimization. Runs vInspect installs end-to-end — from line engineering and on-site commissioning to operator training.

Machine learning · computer vision · software
Dr.-Ing. Christopher Rohkohl
Finament Solutions S.L.
Andorra

Senior machine learning and computer-vision expert. Years of experience developing innovation-driven, high-performance ML and computer-vision systems for medical, industrial and security applications. Designs and builds the vInspect software platform: vinspect_server, the AI defect classifier, multi-camera color calibration and PrintSight analytics.

Engineered in Austria

Mechanics, electronics and on-site integration designed and built in Austria, project-led from Germany.

Software from Andorra

Software developed and supported from Andorra by Finament Solutions S.L.

FAQ

Questions teams ask before they buy.

Book a demo

See vInspect on a label identical to yours.

Send us a reference artwork and a sample of the defects you want caught. We will walk you through a recorded inspection run on a comparable line and scope what an install on yours would look like.

On-site visits and installs across Austria.
Franz Brandl
Single point of contact
Franz Brandl · Virona GmbH

Senior print-industry expert and consultant. Project lead for vInspect installs in Austria — first call for demos, technical questions and commercial enquiries.