Scan4D vs Memo 3D Scanner

Scan4D is a professional-grade reality capture app for iPhone and iPad. It turns your device's LiDAR sensor and camera into a powerful 3D scanner, capturing detailed meshes, depth maps, and camera poses for downstream reconstruction, photogrammetry, and spatial computing. FULL MODE (LiDAR devices) • Real-time mesh reconstruction with live wireframe overlay • 16-bit depth maps captured at adaptive frame rates • Scan capacity metrics: polygon count, drift tracking, session duration • Interactive 3D mesh preview with camera-sampled vertex coloring • Privacy filtering: person segmentation removes humans from meshes, face blurring in RGB exports LITE MODE (all devices) • RGB frame capture with ARKit camera poses • Server-side photogrammetry via exported image bundles • Automatic detection — no configuration needed TIME-SERIES SCANNING • Group scans by Location and re-scan the same space over time • "Extend Scan" loads a red ghost-mesh overlay of previous captures • Adjacent stitching: scan neighboring areas with shared coordinate frames • ARWorldMap relocalization for precise alignment between sessions FLEXIBLE EXPORT FORMATS • Scan4D: Full bundle with metadata, relocalization map, images, depth, and camera data • Polycam: Raw data import compatible with Polycam • RAW: Nerfstudio/COLMAP-compatible transforms.json + images + depth • OBJ / PLY / USDZ: Native mesh formats for MeshLab, Blender, and iOS Quick Look SERVER INTEGRATION • Configure any HTTP(S) upload endpoint in Settings • Direct PUT upload with live progress tracking • Save to Files or AirDrop for offline workflows • Compatible with self-hosted reconstruction pipelines DEVELOPER MODE • Front/back camera switching for testing privacy features • Synthetic IMU, camera, and depth injection for Simulator testing • Vertex mapping diagnostics and debug overlays Designed for researchers, 3D scanning professionals, and spatial computing developers who need raw, high-fidelity sensor data with full control over the reconstruction pipeline. Requires ARKit-capable device. LiDAR-equipped iPhone or iPad Pro recommended for full functionality.

SCANNING Memo exports high-accuracy, COLMAP-compatible capture data, ready for training on your PC with the 3D reconstruction pipeline you prefer. Memo refines ARKit camera poses to improve tracking accuracy and uses diffusion-based depth estimation to generate detailed geometric priors. This reduces reliance on LiDAR and helps 3DGS training converge faster with more stable initial geometry. Memo works on any AR-capable iPhone. On supported Pro models, LiDAR-assisted capture can also be enabled in Settings. ON-DEVICE TRAINING Train 3D Gaussian Splatting scenes entirely on your iPhone. Memo is deeply optimized for Apple hardware, delivering fast, high-quality local reconstruction without sending your captures to external servers. Choose a preset for a quick start, or adjust training parameters for greater control over the final result. Once training is complete, export a camera-path video or scene file directly from your device. You can also upload your scene to SuperSplat with one tap. PRIVACY FIRST Your captures and reconstructions stay on your device. No cloud processing, no mandatory uploads, and no account required.