Unraveling Metameric Dilemma for Spectral Reconstruction: AHigh-Fidelity Approach via Semi-Supervised Learning

Spectral reconstruction from RGB images often suffers from a metameric dilemma, where distinct spectral distributions map to nearly identical RGB values, making them indistinguishable to current models and leading to unreliable reconstructions.