Technology / Solution

Read Beyond
the Visible

Compact optics born from bio-inspired research have evolved into cameras, software and spectral vision.

01 / BIO-INSPIRED ORIGIN

Vision technology
born from bio-inspired research

Micropix’s technology began with bio-inspired lens and camera research at the KAIST Biophotonics Laboratory. Drawing on insect compound eyes, the research explored new ways of capturing images through arrays of small lenses.

The visual structure of Xenos peckii inspired ultrathin camera research using small optical channels.

Results from five bio-inspired imaging papers coauthored by Kyung-Won Jang: wide-angle, high-resolution, fingerprint, depth and HDR, high-speed and low-light imaging
Keum et al. (2018) · Kim et al. (2020) · Jang et al. (2021) · Kim et al. (2022) · Kim et al. (2025)
Micropix camera exploded view showing optics, circuit boards and housing
Micropix camera assembly

02 / MLA CAMERA PLATFORM

Small lenses,
one camera

Building on this research, Micropix develops MLA-based cameras and image processing software. We connect optical design, hardware and image processing to implement the vision solutions our customers’ systems need.

Optical designCamera developmentImage processing

03 / SPECTRAL EXPANSION

Beyond shape and color,
into the spectrum

Spectral vision is one direction for expanding the MLA camera platform. By combining spectral filters with an MLA, we capture wavelength-specific images and analyze spectral differences to reveal clues about composition and condition that RGB images alone may not distinguish.

Spectral filters and 15 microlenses forming wavelength-specific images on a single sensor
Spectral filters → Microlens array → Image sensor → Spectral images
Spectral image capture

Capture multiple spectral bands in a single shot.

Software analysis

Visualize captured data and analyze the target’s spectral characteristics.

04 / FROM SPECTRUM TO CLASSIFICATION

Turning spectral differences
into evidence for classification

Walnut shells and kernels that appear the same color in RGB images can have different spectra. By analyzing these spectral differences and training on the data, we classify shells and kernels in the image.

RGB image

Walnut shells and kernels that appear the same colorOriginal walnut shell and kernel image in the Micropix viewer

Spectral analysis

Analysis of spectral differences between two materialsMeasured walnut shell and kernel spectra from 410 to 950 nm

Classification

Shell and kernel classification after trainingWalnut classification with class0 and class1 bounding boxes
Explore our camerasExplore applications and PoC