Using our proprietary artificial intelligence technology, the Platform can run on commodity hardware to accurately and intelligently provide rich insights and contextual information at the edge - something that hasn’t been possible until now. Other perception solutions just utilize outdated rules-based analytics.
We built everything from scratch, from our databases to our deep learning framework. This means that we have full control over every aspect of the Platform and can deliver customer-centric solutions unlike other offerings that utilize “one-size-fits-all” frameworks and methods.
Our models are trained on data that is actually relevant to customer needs and pain points, not just on the most common training datasets. Pilot’s team is always adding training data for new and existing applications, which helps our models continuously improve.
There are no hardware or firmware changes necessary to use the Platform. Because there is also no need to use the cloud, secure information stays secure and there are no additional costs associated with data transmission and processing.
The Pilot Platform supports most video formats, such as RTSP streams from standard IP cameras.
The Platform utilizes highly accurate, state-of-the-art computer vision models trained for specific applications. These models can be trained to detect general objects, like a person, or very specific objects, like a brown cat. Pilot offers a set of base models; contact us for more information.
Using the Detection results, the Visual Analytics modules generate further insights about the detected objects, such as the age and gender of a detected person.
The Platform outputs information obtained from the Detection and Visual Analytics modules in a format that enables simple integration with any existing systems and tools.
The Platform can run on any device with access to video data, which includes standard IP cameras and smart doorbells.