IdentifyKit

Swift package used to easily integrate classifier coreML models into your code.

Installing with cocoapods

pod 'IdentifyKit'

Quick start

First start by creating a IdentifyKitDelegate, this will handle the result of any identification or failed identification.

extension ViewController: IdentifyKitDelegate {
    func failedToInitialize(error: String) {
        print("Failed to initialize identifier request: \(error)")
    }
    func didIdentifyObject(name: String) {
        print("Identified: \(name)")
    }
    func identifying() {
        print("Identifying")
    }
    func failedToIdentifyObject() {
        print("Identification Failed")
    }
}

Once you have your delegate setup, you can initialize your IdentyKit object. The initializer takes 3 arguments:

• The delegate which we declare above.
• The desired accuracy, which is a float between 0 & 1, will be used to filter out any identifications that are less accurate than this value.
• The model, which can be any image classification model. We’ve used MobileNet in this example.

let classifier = IdentifyKit(delegate: self, accuracy: Configuration.accuracy, model: MobileNet().model)

Once this is done you can make a request:

func identify(image: UIImage) {
    func identify(image: UIImage) {
        let image = UIImage()
        guard let data = image.pngData() else { return }
        classifier.identify(data)
    }
}