Capture results

Every successful session returns one continuous video and a set of frames sampled from the same capture interval.

Result properties

Property	Type	Contents
capturedVideoData	ByteArray?	Complete captured MP4. The device selects a supported codec.
frames	List<ByteArray?>	Sampled JPEG frames; three by default.
faceImage	ByteArray?	Representative JPEG selected from the sampled frames.
faceGeometry	List<MBFaceGeometryModel>	Geometry for the sampled frames.
versionId	String?	Selected constraint version: v1, v2, or v3.
sessionMetadata	String?	Authenticated metadata for the sampled frames.
sessionVideoMetadata	String?	Authenticated metadata for the exact MP4 bytes.

The front-camera preview is mirrored for the user. Captured images and video are stored unmirrored.

Read the result

override fun onSuccess(result: MBCaptureSessionResult?) {
    result ?: return

    val videoBytes = result.capturedVideoData
    val sampledFrames = result.frames
    val recognitionImage = result.faceImage
    val frameGeometry = result.faceGeometry
    val constraintVersion = result.versionId
}

Serialize the biometric request

val biometricRequest = result.serializedData()
// Send biometricRequest to the matching backend endpoint.

serializedData() contains faceImage, frames, and the selected constraint version. The video is exposed separately through capturedVideoData.

Use authenticated session metadata

Set sessionId before adding the predefined fragment, or pass it when starting the custom service:

captureFragment.sessionId = sessionId

// Custom UI:
captureService.startCamera(sessionId)

After capture, send each metadata value with the media it authenticates:

• sessionMetadata with the sampled frame payload
• sessionVideoMetadata with the unchanged capturedVideoData

Do not re-encode or modify the MP4 before using its video metadata because the metadata authenticates the exact captured bytes.