Deep-learning models trained on metalens optical data.
A library of neural networks that recover resolution, correct optical aberrations, enhance low-light frames, and recognise fingerprints and gestures, extending what compact metalens cameras can do.
Our models reconstruct and enhance metalens imagery and pull meaning from it, covering image quality and scene understanding in one toolkit.
SUPER-RESOLUTION IMAGINGMetaOptics' super-resolution module uses deep learning to upscale images captured by metalens cameras, recovering fine detail that conventional interpolation methods cannot reproduce. Trained specifically on metalens optical data, the algorithm reconstructs spatial information beyond the sensor's native resolution — delivering 2× to 4× upscaling with sharp, faithful results. This enables higher-resolution imaging from compact, low-cost metalens camera modules without increasing sensor size. Fingerprint recognition powered by metalens-grade optical resolution 3D depth sensing for accurate, liveness-aware facial and biometric capture Compact sensor integration suited to IoT devices, access control, and mobile platforms
3D Non-Contact RecognitionMetaOptics' 3D non-contact recognition module enables touchless biometric fingerprint detection and identification using metalens camera imagery — no physical contact with a sensor required. A trained neural network detects fingerprint ridge patterns from images captured at working distance, extracts identifying features, and matches them against stored templates. The compact form factor and extended depth-of-field of metalens optics make this ideal for touchless authentication in access control, mobile devices, and secure identification systems. Fingerprint recognition powered by metalens-grade optical resolution 3D depth sensing for accurate, liveness-aware facial and biometric capture Compact sensor integration suited to IoT devices, access control, and mobile platforms
Finger Gesture RecognitionMetaOptics' finger gesture recognition module uses deep learning to detect and classify finger poses from metalens camera feeds in real time. The current model recognizes finger counting gestures from 0 to 9, outputting classification results for downstream application logic. This enables touchless input for smart glasses, IoT devices, and embedded vision systems powered by MetaOptics camera modules. Real-time hand and gesture detection via metalens IoT camera AI model trained for robust detection across varied lighting and backgrounds Compact form factor suited for AR/VR, HUD systems, and smart device integration
Brightening Imaging AIMetaOptics' brightening imaging module uses a trained neural network to enhance brightness, contrast, and visibility in images captured under challenging lighting conditions. Unlike conventional brightness adjustments that uniformly scale pixel values and amplify noise, this AI-driven approach selectively recovers detail in dark regions while preserving highlights and suppressing noise. This is particularly valuable for compact metalens cameras, where the thin, flat optical design constrains aperture size and light collection.
Tuned Image
Image Processed By Our Algorithm
Original High-Res Image
Sharpening ImagingMetaOptics' sharpening imaging module uses deep learning-based deconvolution to restore image clarity in metalens camera output. Metalens optics can introduce wavelength-dependent aberrations — chromatic blur, coma, astigmatism — that reduce sharpness, especially toward the edges of the field of view. A neural network trained on metalens-specific image data reverses these optical degradations in a single inference pass, delivering sharper images with improved contrast across the full field. This effectively extends the useful performance envelope of the metalens design. The underlying technology is patent-pending and received the AAAI 2026 Best Demo Award.
AI Enhanced| # | Module | Key Benefit |
|---|---|---|
| 1 | Super-Resolution Imaging | Higher resolution from compact metalens sensors without increasing hardware cost |
| 2 | 3D Non-Contact Recognition | Touchless biometric fingerprint authentication at working distance |
| 3 | Finger Gesture Recognition | Real-time touchless interaction for smart glasses and IoT devices |
| 4 | Brightening Imaging | Intelligent low-light enhancement that preserves detail and suppresses noise |
| 5 | Sharpening Imaging | Patent-pending, award-winning correction of metalens optical aberrations |
Want to bring metalens processing into your product? Tell us about it.
Contact usConnect with us
Singapore (HQ)
Metaoptics Technologies Pte Ltd. 81 Ayer Rajah Crescent, #01-45 Singapore 139967
United States
Metaoptics Inc. (USA); 1 Ferry Building, Suite 201 San Francisco, CA 94111
