Custom Training
Ouva provides an efficient artifical intelligence platform that is capable of detecting key activities and assets in realtime. Its machine learning pipeline is trained with the most common activities, outfits and equipment found in the patient room.
You may need to track additional activities (e.g. patient pulling on tubes), custom staff outfits (e.g. cleaning crew with unique uniforms), additional assets (e.g. ventilator) and more. To customize Ouva as such, we provide two alternative ways.
You can set up one (simplest) or multiple (most accurate) video cameras to generate video files that can be annotated and made ready to be ingested by Ouva machine learning pipeline. When creating new scenarios, pay attention to following:
- Variety in personal attributes (e.g. gender, skin color, outfit, height) will reduce bias.
- Changing room lighting conditions and peripheral attributes (accessories) with each recording will ensure Ouva can perform well under different environments
We have created a simulation environment that can quickly and realistically recreate all aspects of the patient room, and can generate thousands of images ready to be learned by the Ouva machine learning pipeline within minutes. Get in touch to use this cost and time efficient way to quickly customize Ouva's capabilities.
Last modified 1yr ago