02 Layout
03 Insert
Lumashape
A connected system that turns a phone photo of a tool into a precisely cut foam insert.
Computer Vision: Software that understands what it sees.
Cameras, scans and documents turned into structured information your operation can act on, tested against your real images.
From pixels to a decision, one layer at a time.
One photograph of a conveyor, and the four things software has to do with it. Step through, or lift the layers apart.
Six things software can do with an image.
Most systems combine two or three: detect, then measure; read, then check.
- Object detection
- Finding and locating the things that matter in photos and video, frame by frame.
- Segmentation
- Outlining the exact shape of an object, a defect or a region, pixel by pixel.
- Visual inspection
- Spotting defects, damage and anomalies that are slow or tiring for people to catch.
- Document extraction
- Reading forms, IDs, invoices and scans and turning them into clean, structured data.
- Tracking
- Following objects across frames to count them, time them or understand how they move.
- Visual measurement
- Estimating size, distance, area or fill level from an image, to a stated tolerance.
Will it work on your images?
That is the only question that matters, and it is answered by looking, not by promising. Six things decide it.
Light
Consistent lighting does more for accuracy than a bigger model. Glare, shadow and dusk are tested for, not hoped away.
Angle and distance
What the camera can see decides what the software can know. Sometimes the right fix is moving the camera.
Resolution
A defect has to cover enough pixels to exist. We work back from the smallest thing you need to catch.
Variety
Clean samples prove little. The awkward images, the ones your staff argue about, are the ones that matter.
Volume
A few hundred representative images can prove feasibility. Production accuracy comes from covering the edge cases.
A right answer
Someone has to say what correct looks like. If two experts disagree about an image, the model will too.
A model is one part of four.
Capture
A guided photo flow, a fixed camera or a scanner. Good capture is half the accuracy.
Model
Trained and tested on your images, against a target agreed with the people who rely on it.
Review
Uncertain results go to a person. Their corrections are kept, and the system improves.
Integration
The decision lands in the software your operation already runs. That is where it becomes useful.
What computer vision means here
Computer vision development at SOLOGEN means building software that interprets images and video: detecting and locating objects, segmenting them, inspecting for defects, tracking movement, taking measurements and extracting data from documents, then connecting the result to the systems that act on it.
Who usually needs it
- An operation where people inspect things by eye, all day
- A business typing data out of photos, scans or forms
- A product that needs to recognise what a camera is pointed at
- A team with thousands of images and no way to search them
Where it runs.
On the device
In a phone or tablet app. Instant, works offline, and the image never leaves the hand that took it.
At the edge
On a small computer beside the camera. For production lines and sites with poor connectivity.
In the cloud
For heavier models, large batches and central review. Simplest to update and to scale.
What sits under the model.
Backend
The logic, APIs and integrations behind the interface.
The backend carries your business rules. A well-structured one makes new features cheap and keeps the product stable when usage climbs.
- Node.js
- Fast, event-driven services and APIs that share a language with the frontend.
- NestJS
- Structured, testable Node.js architecture for larger systems and teams.
- .NET
- Enterprise-grade services, especially where they meet existing Microsoft systems.
AI / ML
The parts of a product that read, see, predict and decide.
AI choices affect accuracy, running cost and how much you depend on a single provider. We choose for the task and keep the option to change.
- Python
- The working language of machine learning, data pipelines and model serving.
- AI APIs
- Language and vision models integrated into products and workflows.
- Machine learning models
- Models trained on your own data for prediction, ranking and classification.
- Computer vision
- Detection, inspection and document understanding from images and video.
Cloud & Infrastructure
Where the product runs, scales and recovers.
Infrastructure decides reliability and running cost. We fit the platform to your scale, your team and any systems you already use.
- AWS
- Broad, mature infrastructure for products that need room to scale.
- Azure
- A natural fit for organisations already built around Microsoft.
- GCP
- Strong data and machine-learning services alongside general infrastructure.
- Firebase
- Managed backend services that get mobile and real-time products live quickly.
- Supabase
- A managed PostgreSQL backend with authentication and storage built in.
Before you ask.
How do we know if our problem can be solved with computer vision?
Usually by looking at a sample of your real images. Send a few dozen, including the awkward ones, and a vision specialist will tell you what looks feasible, what would need testing and what would be hard.
How many images do you need?
Less than people expect to start and more than they expect to finish. A few hundred representative images are often enough to prove feasibility. Production accuracy depends on covering the edge cases, which we collect deliberately.
How accurate will it be?
That depends on the images and the task, so we do not quote a figure before testing. What we do is agree a target with the people who rely on the result, based on what a miss actually costs, and measure against it.
Does it run in the cloud or on the device?
Either. On-device and at-the-edge models suit low latency, poor connectivity or sensitive images. Cloud suits heavier models and central review. The choice is made during planning.
Do you build the software around the model as well?
Yes. A model is one part. Our web, mobile and design teams build the capture flow, the review interface and the integrations that turn it into a working system.
How do we get started?
The fastest route is a conversation with a vision specialist. Bring a description of what you need the system to see and, if you can, some example images.
Computer Vision, in the work.
All work