Accelerate AI development
with synthetic data

How it Works

Step 1

Upload CAD-file

Simply upload your 3D model or CAD file to our platform. With support for industry-standard formats, we make it easy to integrate your products into the simulation environment, saving time and minimizing complexity.

Step 3

Varying Lighting Conditions

Set up your inspection scenario by adjusting lighting, camera angles, and other environmental factors. Our intuitive interface simplifies this process, allowing you to create tailored conditions that mimic real inspection environments with ease.

Step 5

Rendering Of Photo-Realistic Images

Generate tens of thousands of photorealistic images at the click of a button. Our platform harnesses powerful rendering engines to produce highly accurate images, enabling seamless AI training without the need for costly real-world data collection.

Step 2

Define Material Characteristics

Effortlessly customize the material properties and textures of your model. Our advanced simulation tools allow you to define realistic surface qualities, ensuring the synthetic data accurately reflects real-world conditions.

Step 4

Distribution of Defect Parameters

Choose defect types and let our system automatically distribute them across your object. With AI-powered defect generation and customizable parameters, the system creates a wide variety of defect scenarios, making your data more robust and scalable.

Step 6

Automatic Label Generation

Labels are automatically generated alongside your images, allowing you to quickly train AI models.

Our system provides an end-to-end solution for synthetic data creation and AI model development, accelerating the process from weeks to days.

Step 1

Upload CAD-file

Step 2

Define Material Characteristics

Step 3

Varying Lighting Conditions

Step 4

Distribution of Defect Parameters

Step 5

Rendering Of Photo-Realistic Images

Step 6

Automatic Label Generation

Our Offer To You

One technology. Two ways to deliver value.

Syn AI

Let Us Use SightHub For You

YOU define the inspection task and the artefacts you want to detect

YOU provide images or CAD-files

WE generate the data, train the model and build the application

YOU get a production-ready application prepared to deliver value

WE work together to manage your application during the full lifecycle

Talk to Sales

Syn Trainer

Take Control & Use SightHub Yourself

WE set up our platform on your infrastructure (optional)

WE onboard and train your team

YOU have full control over your data, models and applications

YOU deploy the application in your production environment

YOU analyze the results and adjust the application if necessary

Get a demo

The SightHub

One platform to mange the whole AI application lifecycle.

Upload, annotate and manage your data.

Upload your data, group it into datasets and define which objects and classes you want to teach your model. You can easily annotate the data for various tasks such as object detection, instance segmentation or classification.

All in one place
Easy to use
Secure

The Sightwise Dashboard Datasets to collect Images

Generate synthetic data from CAD-models.

To make use of our synthetic data engine, upload one or multiple CAD-models, define the defect classes that you expect and where you would expect them to appear on the part geometry. Next, define other relevant inspection parameters such as lighting and texture. Based on your input, our software will prepare the ideal dataset for your task.

Customizable
Expert mode 
Automatic label creation

The Sightwise Dashboard Point Generate to generate synthetic Data

Build complex applications in no time.

Once the training job is done, you can embed your model into an application. You can define input (i.e. one or multiple cameras) and output, as well as all the steps in between, including conditional paths, thresholds and the option to stack multiple models on top of each other.

No-code editor
Multi-model apps
Custom thresholds

The Sightwise Dashboard Applications to Overviev from the Application

Deploy in your production environment.

There is various options for deploying the application in your production environment - from an API endpoint in the cloud to Docker containers and our SightEdge software, which can run locally on your edge device on the line. Thanks to the use of standard interfaces, integration and communication with existing hardware is fairly straight-forward.

Cloud or on-prem
Standard interfaces
Edge computing

The Sightwise Dashboard Deployments to Deploy in your production environment.

Continuous improvements.

Things may change and certain models might need to evolve over time. We give you the tools to adjust your application based on the feedback you receive from the line.

Flexible
Collaborative
Governance

Sightwise Dashboard ML-OPS Model of Feedback Loop

Visualize and analyze your results.

Next-level defect analysis is waiting for you: create a spatially-resolved visualization of your most common defects and monitor trends over time to draw conclusions on predictive maintenance and how to improve your production processes. You can also set up an alerting system in case a certain threshold is met and much more...

Process improvements
Predictive maintenance
Alerting

Sightwise Dashboard from the Analyses, overview from the Analytics

Get a free feasibility check now

Get in touch and we will analyze your case and evaluate,
whether or not our technology will be able to solve your challgene.

Check now

FAQ

Didn't find the answer you were looking for? Please reach out!

How much real-world data do I need to use your product?

None! You can get started with just a CAD-file. However, in case you can get your hands on a few images, we'd recommend uploading them for validation purposes. If you already have a lot of real-world data - great! Let's use that as well and train in hybrid mode.

What if I dont have a CAD-file?

Not a problem. We have the technology in-house to create this for you.

How long does it take to generate a synthetic dataset?

An average synthetic data generation job is taking around 6 hours of computational work - no manual efforts involved.

What kind of materials can be analyzed?

Pretty much anything rigid: metals like steel or aluminium, plastics, rubber, composites, and much more. Are you having a challenge with non-rigid objects? Please get in touch with us and we'll check what's possible.

What is the minimium cycle time that can be achieved by your solution?

Our technology is capable of processing multiple products per second.

Getting in touch.

Let's open up a new chapter of machine vision together. We support leading companies from various industries to manage the entire AI lifecycle. Join us!

+4951176213163