Our technology
Video & Image Recognition
Fine-tuned recognition for screens, scans, inspections and video streams, reaching near-LLM quality at a fraction of the tokens.
Task-specific vision-language models detect, segment and classify what matters in your images and video: controls on an HMI, items on a conveyor, findings on a scan, events on a site camera.
On a customer's 24 screens, a fine-tuned model recognised 498 controls with 6/6 accuracy where the base model scored 4/6. Image data never leaves your site.
Research on arXiv
- An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
- Learning Transferable Visual Models From Natural Language Supervision (CLIP)
- Segment Anything
Used in
See it on your own data
We start with an MVP on your screens, devices or tickets. Request an MVP and bring a real use case.
Request MVP