Create a Unity app that can classify images from the device's camera using our pretrained model

We have used Azure's Custom Vision platform to train our own vision model. The model has been trained using our custom playing cards developed for our game, Regicide. We are building an app in Unity and we need this app to be able to detect a card via the device camera and classify the card based on our pretrained model. Requirements: -The app needs to use the saved model file locally on the device, it cannot use the Custom Vision API as that would require too much data use. -The app needs to be constantly checking the device front camera and sending that data to the classifier code. We don't want uses to have to tap a button to take a picture. -The app needs to build to both Android and iOS devices -The app must have a configurable "confidence" threshold which we can set (eg 80% confidence), if the threshold is met the classifier code should return the matching label We will supply: -The pretrained model and some testing images of the cards that you can print out and use to test your classifier Example of the kind of thing we need: [login to view URL] You might be able to get it working using the method outlined in the video but we weren't able to figure out how to use it with our own model.