AbstractImaging of melanin in the eye is important as the melanin is structurally associated with some ocular diseases, such as age-related macular degeneration. Although optical coherence tomography (OCT) cannot distinguish tissues containing the melanin from other tissues intrinsically, polarization-sensitive OCT (PS-OCT) can detect the melanin through spatial depolarization of the backscattered light from the melanin granules. Entropy is one of the depolarization metrics that can be used to detect malanin granules in PS-OCT and valuable quantitative information on ocular tissue abnormalities can be retrived by correlating entropy with the melanin concentration. In this study, we investigate a relationship between the melanin concentration and some depolarization metrics including the entropy, and show that the entropy is linearly proportional to the melanin concentration in double logarithmic scale when noise bias is corrected for the entropy. In addition, we also confirm that the entropy does not depend on the incident state of polarization using the experimental data, which is one of important attributes that depolarization metrics should have. The dependence on the incident state of polarization is also analyzed for other depolarization metrics.