Medical Imaging
Crop Imaging
Marine Imaging

QCELL’s proprietary Spectral Vision technology changes the way we see the world. Live, multi-channel imaging, spanning both the visible and the non-visible portions of the spectrum, add new dimensions of information.

Spectral Vision is a bio-inspired technology from animals like Mantis shrimps, that have 12 different types of photoreceptor channels. Mantis shrimps can see hues that humans, with just 3 channels, cannot see.

Spectral Vision combines Spectral Artificial Vision with Machine Learning/AI to gather unparalleled analytical power in a live streaming fashion. We can now see better and more! Hidden information becomes visible with the multichannel vision.

Unlike conventional cameras that acquire streams of single images, QCELL’s “Spectral Vision” cameras stream packs of spectral images at video rate. Each spectral cube pack comprise several, full HD narrowband images, an information-rich, multidimensional dataset.

The information contained in the spectral cube is equivalent to a full spectrum per image pixel. Embedded spectral classifiers group all these millions of spectra into color-coded group of pixels, by utilizing efficient spectral similarity metrics. The color-coded classes of pixels make up the so-called spectral map, which depicts comprehensively the dominant spectral profiles in the imaged field.

QCELL’s product pipeline capitalizes on the opentotraining nature of the Spectral Vision platform. Machine learning/AI-based system’s training involves the comparison of spectral profiles corresponding to both target and reference samples (golden standards). This establishes a correlation formula converting the spectral map to either a diagnostic map and/or to a chemical identity map. Diagnostic maps is the last link of imaging chain of QCELL’s Spectral Vision Medical Scopes, whereas chemical identity maps is the output of imaging systems indented for non-destructive testing.

The mantis shrimp (up) and the spectral data cube, showing spectra per pixel and the spectral classification map (first image from the left) (bottom).