Knowledge Base
Deployment Options
The Vigil AI CAID Classifier can be made available in a variety of forms including:
- A standalone Windows service with API
- A Linux Docker container with micro-service with API
- Integrated into a variety of different media analysis tools (contact us for more information)
API Integration
In micro-service form, The Vigil AI CAID Classifier uses a gRPC API which provides access to an image classifier service that is built using state of the art Deep Convolutional Neural Networks. Service communication is over HTTP and the interface is described using the Protocol Buffer interface description language version 3 (Proto3).
By using Proto3, the Classifier service can be accessed easily using a variety of different languages and platforms. For documentation concerning language/platform combinations please see the protocol buffer developer guide.
If you are interested in testing the solution, we are able to provide documentation to demonstrate use of the Classifier service, as well as how to set up a client and how to call some of the available functions.
Licensing Control
Various licensing solutions can be made available depending on the application.
Classifier Operation
The classifier has been trained to categorise images from a set of pre-trained classes. The classify method analyses the contents of images to provide confidence estimates for each pre-trained class. The service can either take the image file names or the file bytes directly.
If you are interested in testing the solution, we are able to provide documentation explaining how to classify images by file name and by file bytes, the optimal processing conditions, and how to interpret various responses.