01
Retail tagging
Products, shelves and store scenes labelled with the attributes your system needs, such as brand, category, facings and placement, so retail models learn to recognise what is on the shelf and where it sits.
Data & annotation
Teach AI to see, with carefully labelled images and video.
An AI model learns to recognise things by studying thousands of examples that people have marked up. We do that marking up: drawing around products, people and objects and naming them, to rules we agree with you first. The work is reviewed before you receive it, and it arrives in the format your team already uses.
Need the raw data collected first?
See web scraping and dataA real example
What a model sees once it has learned from labelled video.
Veesion's software watches a shop's existing cameras and alerts staff to the gestures that tend to come before theft. To do that, it first had to learn from video that people labelled frame by frame. Innovaith labels video like this for Veesion.
Clip and image © Veesion, shown with their permission.
Try it yourself
The kinds of label, one at a time.
Switch between the layers. Each one is a different kind of label we add, so a model can learn to see the same things you do.
Products. A box and a name for every product on show. This teaches a model to find and count what is on the shelf.
Our own illustration, drawn on a stock photo so you can switch each kind of label on and off. It is not client footage.
What we label
Tagging & Annotation, done carefully.
02
Price tagging
Shelf-edge and promotional labels captured and linked to the right product, giving pricing, compliance and promotion analysis a dependable source of data.
03
2D bounding boxes
Tight, consistently drawn rectangles around each object class you define: the standard starting point for object detection and localisation models.
04
3D bounding boxes
Cuboids that capture an object's position, size and orientation in space, for depth-aware work such as robotics, autonomous driving and warehouse inventory.
05
Sports tagging
Players, the ball and key moments tagged across footage, so performance and tactical models learn from data labelled the way analysts think about the game.
What we offer
Also available.
- 01
Polygons & segmentation
Precise outlines, semantic masks and instance segmentation for objects with irregular shapes.
- 02
Keypoints & landmarks
Skeletal keypoints, facial landmarks and product reference points placed consistently across a dataset.
How we work
From guidelines to delivered data.
AI takes the first pass, trained people review it, and quality is checked before delivery.
- 01
Scope
We agree label definitions, edge cases, formats and quality targets with you.
- 02
Pilot
A small batch tests the guidelines before any volume work starts.
- 03
Annotate
Trained annotators label to the agreed guidelines, with tooling to match.
- 04
Review
Multi-stage quality review against the specification, with feedback loops.
- 05
Deliver
Data in your format, with notes on what was checked, then we iterate.
Where it helps
Typical projects.
The kinds of work this service usually supports.
Retail shelf monitoring
Training models that spot empty shelf space, misplaced products and planogram breaks from store photos.
Price compliance checks
Building datasets that let a model read shelf prices and compare them against your price file.
Robotics and autonomous systems
Labelling camera and LiDAR data so vehicles, drones and robots can detect and locate what is around them.
Pose and movement analysis
Keypoint datasets for fitness, sports and ergonomics applications that need to understand body position.
Details
Output formats
- COCO JSON
- YOLO TXT
- Pascal VOC XML
- KITTI
- CVAT XML
- Label Studio JSON
- Segmentation masks (PNG)
- CSV
- Custom JSON schemas
Questions
Can you work with our annotation guidelines?
Yes. We start from your existing guidelines or write a specification with you, then run a pilot batch so both sides can settle edge cases before full production.
How do you check quality?
Batches go through multi-stage review against the agreed specification. We agree quality targets and sampling rules with you before work starts and share review results with each delivery.
Can you annotate inside our own labelling platform?
Yes. We can work in your platform, or in an open-source tool such as CVAT or Label Studio, and deliver in whichever format your pipeline expects.
More data & annotation services
All servicesMatch footage turned into data: who was where, what happened and when, for coaches, scouts and media teams.
Explore service
Software that watches video for you and flags what matters, from recorded footage to live alerts.
Explore service
People who review, rank and correct what an AI writes, so it learns the tone and judgement you want.
Explore service
Idea → Architecture → Prototype → Production → Scale
Talk to us about Tagging & Annotation.
Tell us what you're trying to build, test, automate or improve. We'll tell you honestly where we can make a real difference, and where we can't.