Using OCR

Applies to LUNA ID for Android only.

The OCR (Optical Character Recognition) module enables on-device scanning and recognition of identity and official documents using the device’s built-in camera. It extracts both textual and graphical data (for example, photos, signatures) from supported document types.

Key considerations

Memory usage

The OCR module may use up to 256 MB of RAM during initialization and recognition.

Consider this when:

• Running on devices with limited resources.
• Using OCR concurrently with other resource-intensive modules.
• Launching OCR in the background or within complex screen hierarchies.

Camera permission

The camera permission is required for OCR to function.

Errors

License and LUNA ID errors are always propagated and must be handled at the application level.

Step 1: Add the OCR dependency

To use the OCR functionality, you must explicitly include the OCR module as a dependency in you Android project. To do this, specify it in the build.gradle.kts file:

implementation("ai.company.product:ocr:X.X.X@aar")

Step 2: Activate the OCR license

The OCR functionality requires a valid license entry in the license.conf file:

<?xml version="1.0"?>
<settings>
    <section name="Licensing::Settings">
        <!-- Other license parameters -->
        <param name="OCR" type="Value::String" text="ocrLicense" />
    </section>
</settings>

Important notes:

• If the OCR license parameter is missing, the OCR functionality is disabled.
• If the parameter is present but contains an invalid value, the OCR initialization fails with an error.
• If the parameter is present and valid, OCR is activated successfully.
• OCR licensing is independent of core LUNA ID licensing. Both can be enabled or disabled separately.

Step 3: Initialize OCR

After adding the dependency and confirming the license, initialize the OCR using:

Ocr.initialize(context, licenseFile)

The OCR initialization result is returned via Ocr.ocrInitStateFlow as the OcrInitStatus object. All initialization statuses are defined in the OcrInitStatus class hierarchy:

ealed class OcrInitStatus {
    data object NotInitialized : OcrInitStatus()
    data object InProgress : OcrInitStatus()
    data object Success : OcrInitStatus()
    data object NotIncluded : OcrInitStatus()
    data class Failure(val cause: Throwable? = null) : OcrInitStatus()
}

In case of an error, the corresponding reason will be passed to the flow.

Step 4: Start the OCR

Use the following method to start the OCR:

Ocr.start(context)

After calling the method, a new Activity and the device camera will be launched, and OCR recognition will be started.

Step 5: Handle results

Once the process is completed, the result is sent to LunaID.eventChannel as an event:

sealed class OcrResult {
    data class Success(val data: OcrData) : OcrResult()
    data class Failure(val error: Throwable) : OcrResult()
}

Result	Description
OcrResult.Success	Recognition was successful. Contains the OcrData object, which includes: Document type Text fields Graphical fields (for example, photo, signature)
OcrResult.Failure	Recognition failed. Contains an error object (Throwable) which is an OcrError. All possible OCR errors are defined in the sealed class OcrError.