Characterization of a LED-Based Non-Mydriatic Hyperspectral Retinal Camera

Area of Science:

• Ophthalmic imaging technology
• Medical device engineering
• Spectral imaging systems

Background:

• Advanced retinal diagnostics require high-resolution imaging capabilities.
• Current systems may face limitations in speed, complexity, or non-mydriatic operation.
• Hyperspectral imaging offers potential for detailed tissue characterization.

Purpose of the Study:

• To develop a non-mydriatic hyperspectral retinal camera balancing performance, speed, and simplicity.
• To create a practical system for advanced retinal diagnostics.
• To enable high-content, non-invasive retinal imaging.

Main Methods:

• Integrated LED illumination, linear variable filters, custom optics, and a monochrome sensor.
• Utilized a motorized 3D stage for capturing hyperspectral data across 29 wavebands (450-850 nm).
• Evaluated optical performance and performed in vivo imaging on human subjects for spectral signature capture and repeatability.

Main Results:

• Achieved a 40° field of view with spectral resolution of 20-80 nm.
• Demonstrated chromatic focal correction and uniform illumination across the spectral range.
• Showcased in vivo capture of distinct spectral signatures and high repeatability (spectral variation <5%) with ~300 ms acquisition.

Conclusions:

• The prototype offers reliable, repeatable hyperspectral retinal imaging in a compact, semi-automated design.
• Provides a foundation for enhanced spectral precision, artifact reduction, and expanded field of view.
• Enables non-invasive, high-content retinal imaging suitable for clinical workflows.