This paper introduces a computational framework for recovering spectral information from single-shot photos using a specially designed color chart and algorithm, achieving spectral resolution comparable to scientific spectrometers. It eliminates the need for training data or pre-trained models and has potential applications in accessible optical spectroscopy and hyperspectral imaging.
