Facelandmarks 3D position in IOS Scenekit

[atilagulers]

Hello, i'm trying to convert facelandmarks to 3d points in scenekit.

Even with incorrect depth value face tracks correctly but it is not useful as occluder. because worldpoint is wrong. and face mesh always stays front of every sceneobject or back of every sceneobject.

I've tried to use faceNode.transform = matrix (from mediapipe) it works great but it does not track landmarks.

right now i update geometry vertices by using these landmarks and update geometry uv with uv.json data that mediapipe provided.

How can i get correct world position of landmarks in scenekit?

Thank you.

        if let normalizedLandmarkArray = firstFaceNormalizedLandmarks {
            
            
            let vertexPositionArray = normalizedLandmarkArray.enumerated().map { index, landmark -> simd_float3 in

                let screenX = CGFloat(landmark.x) * imageWidth
                let screenY = CGFloat(landmark.y) * imageHeight
                

                let screenPoint = SCNVector3(screenX, screenY, 0.5)
                
                let worldPoint = sceneView.unprojectPoint(screenPoint)



                return simd_float3(Float(worldPoint.x), Float(worldPoint.y), Float(worldPoint.z))


            }