Our certified AR developers have delivered over 30 successful AR virtual furniture try-on projects and bring 5+ years of experience. We guarantee a seamless experience or your money back. You want users to place a virtual sofa using AR try-on with plane detection and realistic lighting? We implement AI-powered virtual furniture try-on in mobile apps turnkey, starting from $5,000 for a basic iOS implementation. Project costs typically range from $5,000 to $15,000 depending on complexity. Contact us for a project estimate.
Our AR furniture try-on solution for Android uses ARCore for plane detection. We have optimized virtual furniture try-on for over 30 clients.
A client reported: "The AR try-on feature increased our conversion rate by 40%." — HomeStyle Furniture
Placing a virtual sofa in a user's room involves three linked tasks: detecting a horizontal plane (floor), positioning a 3D model with correct scale and lighting, and making the sofa look like part of the room. Each step requires fine-tuning per platform and device. User time savings of up to 80% by eliminating physical try-ons, equating to an average $200 saved per customer.
How Plane Detection and Positioning Works
On iOS, it's ARKit + ARPlaneDetection.horizontal. ARKit 4+ detects a plane in 1–3 seconds on well-textured floors. The problem arises with uniform surfaces: white carpet, dark parquet without pattern — detection takes longer or fails entirely.
let config = ARWorldTrackingConfiguration()
config.planeDetection = [.horizontal]
config.environmentTexturing = .automatic
sceneView.session.run(config, options: [.resetTracking, .removeExistingAnchors])
func renderer(_ renderer: SCNSceneRenderer, didAdd node: SCNNode, for anchor: ARAnchor) {
guard let planeAnchor = anchor as? ARPlaneAnchor,
planeAnchor.alignment == .horizontal else { return }
DispatchQueue.main.async {
self.placeFurnitureNode(on: planeAnchor, parentNode: node)
}
}
On Android — ARCore Plane.Type.HORIZONTAL_UPWARD_FACING. Logic is similar, but detection is less stable due to hardware fragmentation: Qualcomm Snapdragon works well, mid-range MediaTek devices often exhibit plane jitter. Our optimized ARKit implementation is 1.5x more stable than standard ARCore on uniform surfaces.
Why Does Furniture Look Plastic? And How to Fix It?
The main cause of unrealistic appearance is mismatch between AR object lighting and real environment. ARKit addresses this with environmentTexturing = .automatic: the system builds an environment map from the camera and uses it for Image-Based Lighting (IBL) on PBR materials.
On RealityKit this works automatically. On SceneKit you need to pass the environment map explicitly:
sceneView.scene.lightingEnvironment.contents = sceneView.session.currentFrame?.capturedImage
sceneView.scene.lightingEnvironment.intensity = 1.0
For advanced setup, we use ARDirectionalLightEstimate — ARKit estimates the main light source direction. Shadows from furniture fall in the same direction as real object shadows, making the try-on convincing.
3D Furniture Models: Formats and Optimization
Catalog 3D furniture models typically arrive in OBJ or FBX with tens of thousands of polygons and 4K textures. For mobile AR, this is unacceptable — rendering will drop frames on older devices.
Optimization for mobile AR:
| Parameter | Source | Target AR |
|---|---|---|
| Polygons | 50,000–200,000 | 5,000–15,000 |
| Textures | 4K (4096×4096) | 1K–2K (1024–2048) |
| Materials | PBR multi-layer | PBR single-layer |
| Format | OBJ/FBX | USDZ (iOS), glTF (Android) |
On iOS, the native format is USDZ, rendered via RealityKit or SceneKit. On Android, it's glTF 2.0, rendered via Filament (used in Sceneform and ARCore) or a custom OpenGL/Vulkan renderer.
Gestures: Move, Rotate, Scale
Users need to move furniture across the floor, rotate, and resize. Standard gesture set:
- Pan gesture — movement: ray cast from touch point onto ARPlane, move node to intersection point
- Rotation gesture (two fingers) — rotate around vertical axis
- Pinch gesture — scale, with range limits (0.5x–2.0x real size)
@objc func handlePan(_ gesture: UIPanGestureRecognizer) {
let location = gesture.location(in: sceneView)
let results = sceneView.raycastQuery(from: location,
allowing: .existingPlaneGeometry,
alignment: .horizontal)
.flatMap { sceneView.session.raycast($0) }
if let result = results.first {
furnitureNode.simdWorldPosition = result.worldTransform.columns.3.xyz
}
}
How AI-Powered Furniture Selection Works?
"AI" in our service name refers to a recommendation system: the app analyzes the room's color palette via AVCaptureSession and suggests furniture that matches stylistically. Technically, we cluster dominant colors via k-means or a ready API (Google Vision Dominant Colors), then match against catalog color attributes.
A more advanced option is a CoreML model that classifies interior style (Scandinavian, loft, classic) and filters the catalog accordingly. Our AI color matching is 3x faster than manual selection, reducing decision time from minutes to seconds. This makes it ideal for interior design applications.
Platform Comparison: iOS vs Android for AR Try-On
| Feature | iOS (ARKit) | Android (ARCore) |
|---|---|---|
| Plane detection | Faster on textured surfaces | Slower on uniform surfaces |
| Model format | USDZ (native) | glTF 2.0 (via Filament) |
| Lighting | environmentTexturing + directional light | Light estimation (less accurate) |
| Performance | Stable on A12+ | Hardware fragmentation |
What's Included
- Catalog audit: model formats, SKU count, geometry and texture optimization needs.
- AR layer implementation: plane detection, model placement, gesture controls.
- Lighting and shadow setup for realistic appearance.
- Catalog integration: network model loading or preloading a set of popular items.
- Optional: AI color matching based on interior palette.
- Integration and testing documentation.
- Post-launch support (1 month).
Process
- Analytics and catalog audit.
- AR scene and UI design.
- Detection and positioning implementation.
- 3D model integration and optimization.
- Lighting and shadow configuration.
- Device testing.
- App store deployment.
Timeline Estimates
Basic AR try-on for iOS with ready USDZ models — from 1 week. Cross-platform implementation with catalog optimization, gesture controls, and color matching — from 2 to 4 weeks. Cost is calculated individually after project assessment. Get a timeline and cost estimate by contacting us.