On the Equivalence Between Jensen–Shannon Divergence and Michelson Contrast

摘要: This paper focuses on the link between a visible linear and local distortion in pictorial scene its cost terms of information theory quantities. In particular, formal relation Michelson visual contrast Jensen-Shannon divergence (JSD) will be provided. A universal just noticeable threshold is also derived by maximizing JSD constrained to greatest an invisible distortion. Such independent both parameters probability density function degraded image. It then able measure bits budget spent for leaving invisible. Some tests synthetic real distorted images validate theoretical findings. They confirm that proposed really get lower bound human perception it may useful many applications, such as bit allocation coding procedures automatic setting restoration algorithms.