IBM Watson Eğitilmiş Model (Core ML) ile Trafik İşaretleri Tanımlama | iOS --- http://www.bulentsiyah.com/ibm-watson-egitilmis-model-core-ml-ile-trafik-isaretleri-tanimlama-ios/

IBMCustomMeyveTrafikViewController.swift

//
//  IBMCustomMeyveTrafikViewController.swift
//  ML Ornekleri
//
//  Created by Bülent Siyah on 2.07.2018.
//  Copyright © 2018 Bülent Siyah. All rights reserved.
//

import UIKit
import CoreML
import Vision

class IBMCustomMeyveTrafikViewController: UIViewController , UIImagePickerControllerDelegate, UINavigationControllerDelegate {
    
     var meyveTrueTrafikFalse = true
    @IBOutlet weak var image: UIImageView!
    @IBOutlet weak var result: UILabel!
    
    var imagePicker: UIImagePickerController!
    var userPickedImage: UIImage?
    
    override func viewDidLoad() {
        super.viewDidLoad()
        
        imagePicker = UIImagePickerController()
        imagePicker.delegate = self

    }

    func imagePickerController(_ picker: UIImagePickerController, didFinishPickingMediaWithInfo info: [String : Any]) {
        
        if let pickedImage = info[UIImagePickerControllerOriginalImage] as? UIImage {
            image.contentMode = .scaleAspectFit
            image.image = pickedImage
            self.userPickedImage = pickedImage
        }
        
        dismiss(animated: true, completion: nil)
    }
    
    func imagePickerControllerDidCancel(_ picker: UIImagePickerController) {
        dismiss(animated: true, completion: nil)
    }
    
    @IBAction func btnResimSec(_ sender: Any) {
        imagePicker.allowsEditing = false
        imagePicker.sourceType = .photoLibrary
        present(imagePicker, animated: true, completion: nil)
    }
    
    @IBOutlet var btnAnalizEt: UIView!
    
    @IBAction func btnAnalizYap(_ sender: Any) {
        
        do {
            var model : VNCoreMLModel
            if(meyveTrueTrafikFalse){
                model = try VNCoreMLModel(for: ElmaMandalinePortakal().model)
            }else{
                model = try VNCoreMLModel(for: Trafikisaretleri().model)
            }
            
            let request = VNCoreMLRequest(model: model) { (request, error) in
                if let results = request.results as? [VNClassificationObservation], let result = results.first {
                    DispatchQueue.main.async {
                        self.result.text = "\(result.identifier) \(result.confidence * 100)%"
                    }
                }
            }
            request.imageCropAndScaleOption = .centerCrop
            DispatchQueue.global(qos: DispatchQoS.QoSClass.userInitiated).async {
                let handler = VNImageRequestHandler(cgImage: self.userPickedImage!.cgImage!, options: [:])
                do {
                    try handler.perform([request])
                } catch {
                    print(error)
                }
                
            }
        }catch {
            print(error)
        }
    }
    
}