iOS 2.0.x UI only

OVERVIEW

TrustVision SDK is an iOS SDK for TrustVision Engine. It provides these features:

• ID and selfie matching.
• Liveness checking.

Specifications

• Xcode version 10.2+
• target iOS version 9+
• swift version: 5

Integration Steps

1. Adding the SDK to your project

• Drag all file *.framework and *.bundle into project

• Add key to info.plist:

<key>NSCameraUsageDescription</key>
<string>Open camera</string>

• Add dependencies
  • Use CocoaPods, add these lines to podfile

pod 'TensorFlowLiteSwift', '~> 2.4.0'
pod 'PromiseKit', '~> 6.8'
pod 'CocoaLumberjack/Swift'

Those lines are added at the end of podfile

  post_install do |installer|
    installer.pods_project.targets.each do |target|
      target.build_configurations.each do |config|
        config.build_settings['BUILD_LIBRARY_FOR_DISTRIBUTION'] = 'YES'
      end
    end
  end

• Add manually:
  • Add these frameworks and chose Embed & Sign
    • CocoaLumberjack.framework
    • PromiseKit.framework
  • Add these frameworks and chose Do Not Embed
    • TensorFlowLite.framework
    • TensorFlowLiteC.framework
    • Add key Validate Workspace in Build Settings
• If host app is using Objective-C, please follow these additional steps:
  • add a empty swift file and create bridging file (to force project create key SWIFT_VERSION)
  • add flag -lc++ and -ObjC in Other linker flags

2. Start the SDK

The SDK provides some built in Activities example activity to capture id, selfie, liveness...

2.0. Initialize SDK

Initialize the sdk

  TrustVisionSdk.shared.initialize(clientSettingsJsonString: nil, languageCode: "vi")

Options:

• clientSettingsJsonString: String. The jsonConfigurationByServer is optional but recommended. It's the setting specialized for each client from TS server. It's the response json string get by API https://ekyc.trustingsocial.com/api-reference/customer-api/#get-client-settings. When it's null or unmatched with the expected type then the default setting in the SDK will be used.
• languageCode: String. Language code. vi or en

2.1. Update and get sdk language

Allow user to change the sdk language after initialization

TrustVisionSdk.shared.changeLanguageCode(languageCode: String)
TrustVisionSdk.shared.getLanguageCode() -> String?
TrustVisionSdk.shared.getSupportedLanguageCodes() -> []

2.2. Capture the ID

The id capturing activity will show the camera to capture image, preview the image. To start the id capturing activity.

2.2.1. Set config parameters

let config = TVIdCardConfiguration(
                cardType: selectedCardType,
                cardSide: TVIdCardConfiguration.TVCardSide.front,
                isSoundEnable: false,
                isReadBothSide: false,
                skipConfirmScreen: false,
                isEnablePhotoGalleryPicker: false)

Options:

• cardType: TVCardType. List of supported cards can be found by let cardTypes: [TSCardType] = try TrustVisionSdk.getCardTypes()
• cardSide: TVCardSide. Card side to capture
• isEnableSound: Bool. Sound should be played or not
• isReadBothSide: Bool. If true then the sdk will capture both side if possible; otherwise, then the card side defined in cardSide will be used
• skipConfirmScreen: Bool. Skip id capturing confirmation screen nor not
• isEnablePhotoGalleryPicker: Bool. Allow user select id card image from phone gallery

2.2.2. Start id capturing from configuration

let vc = TrustVisionSdk.shared.startIdCapturing(configuration: config, success: { (result) in

}, failure: { (error) in

}, cancellation: {
    // sdk is canceled
})

where:

• configuration: TVIdCardConfiguration
• success method that will be called in case success. Parameters:
  • result : TVDetectionResult. Use the following fields:
    • frontIdImage
    • backIdImage
• failure: FailureCallback
• cancellation: CancellationCallback

2.2.3. Handle ID capturing results

2.2.3.1. Get Image Ids to be used in a particular use case

Use this API https://ekyc.trustingsocial.com/api-reference/customer-api/#upload-image The images should be uploaded as JPEG data with 100% quality. For example:

// with front side
let dataToUpload = result.frontIdImage.rawImage.jpegData(compressionQuality: 1.0) // 100% quality

// with back side
let dataToUpload = result.backIdImage.rawImage.jpegData(compressionQuality: 1.0) // 100% quality

2.2.3.2. Upload QR images

if result.frontIdQr.isRequired is true then result.frontIdQr.images array should be non-empty. Otherwise, clients should be warned to re-capture id card photos.

QR imagegs will be uploaded with this api: https://ekyc.trustingsocial.com/api-reference/customer-api/#upload-image

• Fields:
  • data: result.frontIdQr.images[i].raw_image_base64
  • label: result.frontIdQr.images[i].label
  • metadata: result.frontIdQr.images[i].metadata

*The same logic will be applied to result.backIdQr

2.3. Capture the selfie

The selfie capturing activity will show the camera to capture image, preview the image and verify active liveness in local. To start the selfie capturing activity.

2.3.1. Set config parameters


let config = TVSelfieConfiguration(
                cameraOption: TVCameraOption.front,
                isSoundEnable: true,
                livenessMode: self.config.livenessMode,
                skipConfirmScreen: false)

Options:

• cameraOption: TVCameraOption. Set the camera mode
• isSoundEnable: Bool. Sound should be played or not
• livenessMode: TVLivenessMode. Set the liveness verification mode
• skipConfirmScreen: Bool. Skip selfie capturing confirmation screen or not

2.3.2. Start selfie capturing from configuration

let vc = TrustVisionSdk.shared.startSelfieCapturing(configuration: config, success: { (result) in

}, failure: { (error) in

}, cancellation: {
    // sdk is canceled
})

where:

• configuration: TVSelfieConfiguration
• success method that will be called in case success. Parameters:
  • result : TVDetectionResult . Use the following fields:
    • selfieImages
• failure: FailureCallback
• cancellation: CancellationCallback

2.3.3. Handle selfie image list

1. Use this api to get image id: https://ekyc.trustingsocial.com/api-reference/customer-api/#upload-image The images should be uploaded as JPEG data with 100% quality. For example:

// with frontal images
let dataToUpload = result.selfieImages[i].frontalImage.rawImage.jpegData(compressionQuality: 1.0) // 100% quality

// with gesture images
let dataToUpload = result.selfieImages[i].gestureImage.rawImage.jpegData(compressionQuality: 1.0) // 100% quality

id of frontal image i = image id of result.selfieImages[i].frontalImage.rawImage id of gesture image i = image id of result.selfieImages[i].gestureImage.rawImage

2. Use this api to get video id: https://ekyc.trustingsocial.com/api-reference/customer-api/#upload-videoaudioframes id of selfie video i = video id of result.livenessVideos[i]

3. Call this api to check liveness https://ekyc.trustingsocial.com/api-reference/customer-api/#verify-face-liveness with params:

In images field, each element contains:

{
  "id": "<id of frontal image i>"
}

In gesture_images field, each element contains:

{
  "gesture": "lower case string of <result.selfieImages[i].gestureType.description>",
  "images": [
    {
      "id": "<id of gesture image i>"
    }
  ]
}

videos field is a list which comprises 2 lists:

• (i) Liveness videos. Each element contains:

{
    "id": "<id of selfie video i>" (step 2)
}

• (ii) Liveness frames list:

result.getSelfieFrameLists() // func getSelfieFrameLists() -> [[String: Any]]

With metadata field is result.livenessMetadata

API references

1. TVDetectionResult

2. TVGestureImage

• gestureType: GestureType
• frontalImage: TVEncryptedImage
• gestureImage: TVEncryptedImage

3. TVGestureImage.GestureType

• up
• down
• left
• ight
• frontal

4. TVCameraOption (Enum)

• TVCameraOption.front: Use front camera
• TVCameraOption.back: Use back camera
• TVCameraOption.both: The screen will have a button to switch between front & back camera

5. TVLivenessMode (Enum)

• TVLivenessMode.none: no liveness verification. Just capture the selfie and verify sanity.
• TVLivenessMode.passive: Use texture-based approach.
• TVLivenessMode.active: Use challenge-response approach. User needs to follow and finish all steps when capturing selfie like turn left, right, up, smile, open mouth...

7. TVCardQr

• isRequired: This side of card contains QR or not
• images: [TVEncryptedImage]. Array of QR images

6. TVLivenessResult

• isLive: selfie is live or not

8. FailureCallback (Callback)

Will be called in case failed. Parameters:

• error: TVError.
  • errorCode: the specific error code.
  • description: the human-readable error description can be show to end user

9. CancellationCallback (Callback)

Will be called in case the sdk is cancelled. No parameters