How to use AWS Rekognition to Compare Face in Swift 3

Question

I've been trying to use the AWSRekognition SDK in order to compare face. However, Amazon has no Documentation on how to integrate their SDK with iOS. They have links that show how to work with Recognition (Developer Guide) with examples only in Java and very limited.

I wanted to know if anyone knows how to integrate AWS Rekognition in Swift 3. How to Initialize it and make a request with an image, receiving a response with the labels.

I have AWS Signatures AccessKey, SecretKey, AWS Region, Service Name. also Body

{
  "SourceImage": {
    "S3Object": {
      "Bucket": "bucketName",
      "Name": "ios/sample.jpg"
    }
  },
  "TargetImage": {
    "S3Object": {
      "Bucket": "buketName",
      "Name": "ios/target.JPG"
    }
  }
}

how can I initialize Rekognition and build a Request.

Thanks you!

Answer 1

1. Instantiate the Rekognition Client, Here I'm using the client with the default configuration.

   let rekognitionClient:AWSRekognition = AWSRekognition.default()
   

Otherwise, you can use the credentials as follows:

    let credentialsProvider = AWSCognitoCredentialsProvider(
        regionType: AWSRegionType.usEast2,
        identityPoolId: "us-east-2_myPoolID")

    let configuration = AWSServiceConfiguration(
        region: AWSRegionType.usEast2,
        credentialsProvider: credentialsProvider)

    AWSServiceManager.default().defaultServiceConfiguration = configuration
    let rekognitionClient:AWSRekognition = AWSRekognition.default()

2. Now construct the request and set the image in it.

   let image = UIImage(named: "MyImage")
   let request = AWSRekognitionDetectLabelsRequest() 
   request.image = image
   request.maxLabels = <num_labels_needed>
   request.minConfidence = <confidence_interval_needed>
   

3. Now to compare faces, read about the CompareFacesRequest: https://github.com/aws/aws-sdk-ios/blob/master/AWSRekognition/AWSRekognitionService.m#L288

There is a sample test in the SDK that compares two faces in ObjC but you can translate that in Swift:

https://github.com/aws/aws-sdk-ios/blob/master/AWSRekognitionUnitTests/AWSGeneralRekognitionTests.m#L60

    let key = "testCompareFaces"
    let configuration = AWSServiceConfiguration(region: AWSRegionUSEast2, credentialsProvider: nil)
    AWSRekognition.register(with: configuration, forKey: key)
    AWSRekognition(for: key).compareFaces(AWSRekognitionCompareFacesRequest()).continue(withBlock: {(_ task: AWSTask) -> Any in
        print("completed")

Answer 2

Swift 5.0

let key = "testCompareFaces"
let credentialsProvider = AWSStaticCredentialsProvider(accessKey: "Add_access_key_id", secretKey:"Add_secret_access_key_id")
let configuration = AWSServiceConfiguration(region:.USEast1, credentialsProvider:credentialsProvider)
AWSServiceManager.default().defaultServiceConfiguration = configuration
    
AWSRekognition.register(with: configuration!, forKey: key)
let rekognition = AWSRekognition(forKey: key)

guard let request = AWSRekognitionCompareFacesRequest() else {
    puts("Unable to initialize AWSRekognitionDetectLabelsRequest.")
    return
}
let sourceImage = AWSRekognitionImage()
sourceImage!.bytes = sourceImage.jpegData(compressionQuality: 0.4)// Specify your source image
request.sourceImage = sourceImage

let targetImage = AWSRekognitionImage()
targetImage!.bytes = targetImage.jpegData(compressionQuality: 0.4) // Specify your target image
request.targetImage = targetImage

rekognition.compareFaces(request) { (respone, error) in
    if error == nil {
        if let response = respone {
            if let first = response.faceMatches?.first {
                print(first)
            }
        }
    } else {
        print(error?.localizedDescription)
    }
}