MNC Identifier SDK

MNC Identifier is a service to identify, and verify consumer with AI in it.

Table

Liveness Detection

Liveness Detection using mlkit face recognition to detect live person present at the point of capture.

Requirements

Min SDK 21

Setup

build.gradle (root)

repositories {
    ...
    maven { url 'https://jitpack.io' }
}

build.gradle (app)

dependencies {
    implementation "com.github.mncinnovation.mnc-identifiersdk-android:core:1.0.1"
    implementation "com.github.mncinnovation.mnc-identifiersdk-android:face-detection:1.0.1"
}

AndroidManifest.xml

<application ...>...<meta-data android:name="com.google.mlkit.vision.DEPENDENCIES"
android:value="face" /></application>

How To Use

Start liveness activity

startActivityForResult(MNCIdentifier.getLivenessIntent(this), LIVENESS_DETECTION_REQUEST_CODE)

companion object {
    const val LIVENESS_DETECTION_REQUEST_CODE = xxxx
}

Get Liveness Result

override fun onActivityResult(requestCode: Int, resultCode: Int, data: Intent?) {
    super.onActivityResult(requestCode, resultCode, data)
    if (resultCode == RESULT_OK) {
        when (requestCode) {
            LIVENESS_DETECTION_REQUEST_CODE -> {
                //get liveness result
                val livenessResult = MNCIdentifier.getLivenessResult(data)
                livenessResult?.let { result ->
                    if (result.isSuccess) {  // check if liveness detection success
                        // get image result
                        val bitmap = result.getBitmap(this, DetectionMode.SMILE)
                    } else {  //Liveness Detection Error
                        //get Error Message
                        val errorMessage = result.errorMessage
                        //get Error Type (OOM / Exception)
                        val errorType = result.errorType
                    }
                }
            }
        }
    }
}

Customize Detection Sequence

Default detection sequence is HOLD_STILL > BLINK > OPEN_MOUTH > SHAKE_HEAD > SMILE. You can cutomize detection sequence using following method

//the first boolean value indicates if the given detection sequence should be shuffled.
MNCIdentifier.setDetectionModeSequence(
    false, listOf(
        DetectionMode.HOLD_STILL,
        DetectionMode.BLINK,
        DetectionMode.OPEN_MOUTH,
        DetectionMode.SMILE,
        DetectionMode.SHAKE_HEAD
    )
)

Low Memory Threshold

If the memory stack allocation usage is lower than 50 mb (by default), a popup warning will appear before executing the image capture or object detection process. However, you can customize the threshold or you can also disable it by setting it to 0.

MNCIdentifier.setLowMemoryThreshold(50) // for face detection

OCR (Optical Character Recognition)

Optical Character Recognition using mlkit text recognition to detect text at the point of capture.

Requirements

Min SDK 21

Setup

build.gradle (root)

repositories {
    ...
    maven { url 'https://jitpack.io' }
}

build.gradle (app)

dependencies {
    implementation "com.github.mncinnovation.mnc-identifiersdk-android:core:1.1.0"
    implementation "com.github.mncinnovation.mnc-identifiersdk-android:ocr:1.1.0"
}

AndroidManifest.xml

    <application ...>    
        ...    
        <meta-data android:name="com.google.mlkit.vision.DEPENDENCIES"
                    android:value="ocr" />
    </application>

If you use face and ocr, AndroidManifest.xml

    <application ...>    
        ...    
        <meta-data android:name="com.google.mlkit.vision.DEPENDENCIES"
                    android:value="face, ocr" />
        </application>

How To Use

Get OCR result by using scan camera from MNC Identifier

Optional configuration to show button flash at camera activity and show camera only screen.

Default value withFlash is false.
Default value cameraOnly is false.
Default value lowMemoryThreshold is 50.

//call this function before startCapture
MNCIdentifierOCR.config(withFlash = true, cameraOnly = true, lowMemoryThreshold = 70)

Start to capture OCR result activity

//start directly
MNCIdentifierOCR.startCapture(this@MainActivity)

//or with requestCode value
MNCIdentifierOCR.startCapture(this@MainActivity, CAPTURE_EKTP_REQUEST_CODE)

companion object {
    const val CAPTURE_EKTP_REQUEST_CODE = xxxx
}

Get Capture OCR Result

override fun onActivityResult(requestCode: Int, resultCode: Int, data: Intent?) {
    super.onActivityResult(requestCode, resultCode, data)
    if (resultCode == RESULT_OK) {
        when (requestCode) {
            CAPTURE_EKTP_REQUEST_CODE -> {
                val captureOCRResult = MNCIdentifierOCR.getOCRResult(data)
                captureOCRResult?.let { result ->
                    if (result.isSuccess) {
                        result.getBitmapImage(this)?.let {
                            //get image result
                            binding.ivKtpCapture.setImageBitmap(it)
                        }
                        //show all of data result
                        binding.tvCaptureKtp.text = result.ktpModel.toString()
                    } else {
                        //get Error Message
                        val errorMessage = result.errorMessage
                        //get Error Type (OOM / Exception)
                        val errorType = result.errorType
                    }
                }

            }
        }
    }
}

//another option (using registerForActivityResult) 
private val resultLauncherOcr =
    registerForActivityResult(ActivityResultContracts.StartActivityForResult()) { result ->
        if (result.resultCode == Activity.RESULT_OK) {
            val data = result.data
            val captureOCRResult = MNCIdentifierOCR.getOCRResult(data)
            captureOCRResult?.let { ocrResult ->
                if (ocrResult.isSuccess) {
                    ocrResult.getBitmapImage(this)?.let {
                        binding.ivKtp.setImageBitmap(it)
                    }
                    binding.tvScanKtp.text = captureOCRResult.toString()
                } else {
                    //get Error Message
                    val errorMessage = ocrResult.errorMessage
                    //get Error Type (OOM / Exception)
                    val errorType = ocrResult.errorType
                }
            }
        }
    }

MNCIdentifierOCR.startCapture(this@MainActivity, resultLauncherOcr)

Get OCR result by using function extract data. MNCIdentifier will receive input image from your app.

    //call this function to get extract data from uri of image 
    MNCIdentifierOCR.extractDataFromUri(
        uriList,
        this@MainActivity,
        object : ExtractDataOCRListener {
            override fun onStart() {
                Log.d("TAGAPP", "onStart Process Extract")
            }
    
            override fun onFinish(result: OCRResultModel) {
    
                result.getBitmapImage()?.let { bitmap ->
                    binding.ivKtp.setImageBitmap(bitmap)
                }
                binding.tvScanKtp.text = result.toString()
    
            }

            override fun onError(message: String?, errorType: ResultErrorType?) {
                //handle on error here
            }
        })

Extract data OCR input options:

Uri : uri image file
List : list of uri image files
String : path image file
List : list of path image files

mncinnovation / mnc-identifiersdk-android Goto Github PK