Welcome to Eydara
Eydara is an AI-based vision platform deployed on-premise at inspection sites. It brings together several stages into a single pipeline — from raw image capture to a deployed, continuously improving vision model.
The pipeline at a glance
Section titled “The pipeline at a glance”| Stage | What it does |
|---|---|
| Capture | Cameras at inspection rigs photograph products as they pass through. |
| Preprocess | Raw images are adjusted — corrected for lens distortion, cropped, or enhanced — before analysis. |
| Ingest | Processed images are stored and registered so the rest of the system can access them. |
| Visualize and annotate | Domain experts review images, see model predictions, and provide annotations that teach the model. |
| Train | Annotations are used to train or refine vision models. |
| Deploy | Updated models are deployed back to the inspection site, closing the loop. |
Who this documentation is for
Section titled “Who this documentation is for”This documentation is written for the domain expert — the person who knows the product and the production process, not the person who builds or trains the model.
In the graphical interface, you can:
- Browse images captured by cameras at inspection rigs
- Review what the AI model predicted for each image
- Annotate images with your expert knowledge
- Track model performance over time through statistics and threshold tuning
The annotations you provide are what make the model better. Your knowledge of the product is the foundation the system builds on.
What’s next?
Section titled “What’s next?” Home Browse and select images from the main view.
Filtering Learn how to narrow down the image grid using filters.
Thresholds Configure confidence thresholds for model predictions.
Statistics Review prediction and annotation distributions over time.
Runs Manage and compare model comparison runs.
How-To Guides Step-by-step workflows for common tasks.
Glossary Definitions of key terms used in Eydara.