Technology

From algorithm alignment to physical alignment.

iEYELAB's common-path architecture lets different sensing modalities share the same optical origin, dramatically reducing registration effort and system complexity.

One optical path · three sensors

Shared optical axis

All sensing channels originate from the same viewpoint, creating natural pixel-level alignment.

Recombined sensor stack

RGB, depth, polarization and spectrum can be flexibly recombined for new applications.

Lower downstream compute

Physical alignment removes large chunks of pre-processing, freeing inference power upstream.

Expanded sensing dimensions

RGB

Color, texture and appearance.

Depth

3D spatial structure and geometric distance.

Polarization

Material surface, reflectivity and transparent object recognition.

Spectrum

Material composition and wavelength signatures.

Hardware-level multi-modal algorithm stack

16-channel spectrum / iToF phase / polarization RAW parallel decoding

Early fusion on physically aligned multi-modal data

Segmentation and multimodal reasoning on top of a coherent visual base

Structured cognitive graph output for machines to act on

Why the architecture matters

Traditional

iEYELAB

Depth capability

Often requires extra LiDAR, structured light or additional devices.

Native module-level integration without external helper sensors.

Spectral identification

Limited channels and poor synchronization across stacked cameras.

Built-in high-sensitivity multi-spectral channels with full-frame synchronized capture.

Polarization perception

Usually needs external scientific cameras and heavy computation.

Native four-direction polarization imaging with direct hardware output.

Alignment quality

Separated devices create unstable parallax and drifting registration error.

Same-source, same-viewpoint data with physical coaxial alignment.

IP and defensibility

Single-lens heterogeneous common-path sensing system

Zoom-capable common-path sensing system

3D reconstruction for highly reflective objects using polarization and iToF

Joint calibration and solving for spectral and spatial geometry features

Common-path fusion algorithm for polarization, spectrum and depth

3000+

Academic citations

50+

Top-tier papers

Engineering specification targets

FOV

D:90° × H:70°

Working distance

0.3m ~ 5m

ToF wavelength

940nm ± 30nm

Spectral wavelength

400 ~ 700nm

Latency

< 10 ms

Positioning precision

< 1%