System gamma as a function of image- and monitor-dynamic range

Abstract

System gamma is the end-to-end exponent that describes the relationship between the relative luminance values at capture and the reproduced image. The system gamma preferred by subjects is known to vary with the background luminance condition and the image in question. We confirm the previous two findings using an image database with both high and low dynamic range images (from 102 to 107), but also find that the preferred system gamma varies with the dynamic range of the monitor (CRT, LCD, or OLED). We find that the preferred system gamma can be predicted in all conditions and for all images by a simple model that searches for the value that best flattens the lightness distribution, where lightness is modeled as a power law of onscreen luminance. To account for the data, the exponent must vary with the viewing conditions. The method presented allows the inference of lightness perception in natural scenes without direct measurement and makes testable predictions for how lightness perception varies with the viewing condition and the distribution of luminance values in a scene. The data from this paper has been made available online.

Derivatives and inverse of cascaded linear+nonlinear neural models

“…Note that the bimodal PDFs of x 3 and x 4 are consistent with the predictive behavior reported in [80]. Equalization behavior at x 1 reported at [69] is not that evident in this example because this is a low-dynamic-range image.…”

“…5 and 10 (as suggested in [1]), here we also consider particular alternatives for the nonlinearities that have been proposed to account for the response at specific stages in the visual pathway. Namely, the Wilson-Cowan equations [2,3], which could account for the masking between local-oriented sensors [29]; and nonlinear models of brightness perception such as the ones used in tone mapping [4,68,69]. The consideration of these alternatives for specific stages stresses the generality of the proposed framework since, as shown in the examples of the Discussion, the network equations can be applied no matter the specific functional form of each stage (provided the elementary derivatives and inverses are known).…”

“…The transition of γ between γL and γH happens around the value |y| = µ1. This transition is smooth, and its sharpness is controlled by the exponent m. This expression for γ has statistical grounds since the resulting nonlinearity approximately equalizes the probability density function (PDF) of luminance values in natural scenes [69,71], which is a sensible goal in the information maximization context [72]. This nonlinearity can be applied both to linear luminance values [68,73] as well as to linear opponent color channels [46,74].…”

The Role of Lightness Perception in Determining the Perceived Contrast of Real World Scenes

“…The logical extension of our model would suggest that a binary stimulus, composed of zeros and ones (maximum and minimum luminance values) would receive the highest image quality scores as they maximize the standard deviation of a stimulus. This finding it at loggerheads with other finding on image quality, in particular, the finding that image quality scores may be related to the degree of histogram equalization in a stimulus [12]. Stimuli that maximize the standard deviation include a salt and pepper noise, or a simple step function.…”

Improved High Dynamic Range Video Coding with a Nonlinearity Based on Natural Image Statistics

“…The first stage of [9] is developed according the theory developed in [10] and a model of the statistics of natural images [11]. Although the original approach produces tone curves that are image-dependent, instead, we have chosen a fixed and image-independent non-linear function that works well on various sequences for video coding.…”