®REBOTNIX® / Technology
Tools and methods behind our systems.
Here we show how REBOTNIX® trains models, runs them in deployment and processes image data. These tools and methods provide the technical foundation for KINEVA®, HORIZON and our hardware.
KINEVA® TRAINING SUITE
Training models for their task.
The KINEVA® Training Suite is our in-house software for training visual models. We have been building the training methods and tools since 2014. The demo shows synthetic image data from Unreal Engine.
KINEVA® INFERENCE ENGINE
Running models on edge devices and servers.
The KINEVA® Inference Engine is the runtime software for our models. It processes new data on embedded edge devices or x86 servers. REBOTNIX® optimises models, engine and hardware together for each application.
CONTEXT AI COMPRESSION
Using image data with less transmission overhead.
Context AI Compression is an image compression method. It aims to reduce data volume while preserving relevant content such as text and structured details. Our demo illustrates how it works and its technical limits.
Explore the method and demo ↗