Transferring the World’s True Colors to Screens: A Retina-Inspired Multispectral Approach

Current imaging and display systems are built around engineering simplifications of human color perception, typically using only three channels: red, green, and blue (RGB). This model reduces the complexity and realism of natural colors, compressing the vast color gamut perceived by the human eye into a limited color space. Our proposed technology bridges this gap by mimicking the neural color-coding mechanisms of the human retina.

This presentation introduces a neural model for the demultiplexing of signals from the three retinal photodetectors (L, M, S) into seven perceptual color channels: red, yellow, green, cyan, blue, purple, and white. This biologically inspired model can be implemented in an electronic module to derive color channel outputs from a RGB camera, compatible with display systems using multispectral subpixels, such as RGBYC layouts. The approach enables a more natural and vivid reproduction of colors, closely reflecting how they appear in real-world environments.

Diaconu also presents an algorithm for transforming RGB signals into RGBYC format, suitable for modern displays with extended color primarily. Applications range from medical imaging and digital restoration to virtual reality and high-fidelity television. This approach has the potential to redefine how digital systems interpret and display color—with fidelity to human vision.

About the presenter
Vasile Diaconu is a physicist and associate professor at the Université de Montréal, specializing in retaining inspired imaging systems. With more than 30 years of experience in physiological optics, color perception and biomedical signal processing, he holds several patents and is the author of numerous publications in color vision and multispectral analysis. He is also the inventor of techniques for demultiplexing color channels in retinal structures, bridging neuroscience and engineering in the field of color imaging.