Gabriele Simone

Measuring perceptual contrast in digital images

In this paper we present a novel method to measure perceptual contrast in digital images. We start from a previous measure of contrast developed by Rizzi et al. [26], which presents a multilevel analysis. In the first part of the work the study is aimed mainly at investigating the contribution of the chromatic channels and whether a more complex neighborhood calculation can improve this previous measure of contrast. Following this, we analyze in detail the contribution of each level developing a weighted multilevel framework. Finally, we perform an investigation of Regions-of-Interest in combination with our measure of contrast. In order to evaluate the performance of our approach, we have carried out a psychophysical experiment in a controlled environment and performed extensive statistical tests. Results show an improvement in correlation between measured contrast and observers perceived contrast when the variance of the three color channels separately is used as weighting parameters for local contrast maps. Using Regions-of-Interest as weighting maps does not improve the ability of contrast measures to predict perceived contrast in digital images. This suggests that Regions-of-Interest cannot be used to improve contrast measures, as contrast is an intrinsic factor and it is judged by the global impression of the image. This indicates that further work on contrast measures should account for the global impression of the image while preserving the local information.

Termite Retinex: a new implementation based on a colony of intelligent agents

Abstract. The original presentation of Retinex, a spatial color correction and image enhancement algorithm modeling the human vision system, as proposed by Land and McCann in 1964, uses paths to explore the image in search of a local reference white point. The interesting results of this algorithm have led to the development of many versions of Retinex. They follow the same principle but differ in the way they explore the image, with, for example, random paths, random samples, convolution masks, and variational formulations. We propose an alternative way to explore local properties of Retinex, replacing random paths by traces of a specialized swarm of termites. In presenting the spatial characteristics of the proposed method, we discuss differences in path exploration with other Retinex implementations. Experiments, results, and comparisons are presented to test the efficacy of the proposed Retinex implementation.

Measuring perceptual contrast in a multi-level framework

In this paper, we propose and discuss some approaches for measuring perceptual contrast in digital images. We start from previous algorithms by implementing different local measures of contrast and a parameterized way to recombine local contrast maps and color channels. We propose the idea of recombining the local contrast maps and the channels using particular measures taken from the image itself as weighting parameters. Exhaustive tests and results are presented and discussed, in particular we compare the performance of each algorithm in relation to perceived contrast by observers. Current results show an improvement in correlation between contrast measures and observers perceived contrast when the variance of the three color channels separately is used as weighting parameter for local contrast maps.

Measuring perceptual contrast in uncontrolled environments

Contrast is one of the most relevant perceptual and quality factors in digital images and measuring it is not a trivial task. We have carried out an online psychophysical experiment to register perceived contrast. The results from the observers indicate that color images are rated higher than their respective greyscale one indicating that contrast is influenced color. A two-sided sign test at 5% level confirms this hypothesis. A comparison with a lab controlled environment experiment has been carried out in order to investigate possible differences. A statistical analysis of the two experiments indicate that the mean ratings of the observers are not significantly different. A decrease in correlation of previously developed contrast measures can be noticed and a comparison of the correlation coefficients indicate that measuring contrast in uncontrolled environments can be significantly different than measuring in a lab controlled environment.

Multi-level contrast filtering in image difference metrics

In this paper, we present a new metric to estimate the perceived difference in contrast between an original image and a reproduction. This metric, named weighted-level framework ΔEE (WLF-DEE), implements a multilevel filtering based on the difference of Gaussians model proposed by Tadmor and Tolhurst (2000) and the new Euclidean color difference formula in log-compressed OSA-UCS space proposed by Oleari et al. (2009). Extensive tests and analysis are presented on four different categories belonging to the well-known Tampere Image Database and on two databases developed at our institution, providing different distortions directly related to color and contrast. Comparisons in performance with other state-of-the-art metrics are also pointed out. Results promote WLF-DEE as a new stable metric for estimating the perceived magnitude of contrast between an original and a reproduction.

On the Use of Gaze Information and Saliency Maps for Measuring Perceptual Contrast

In this paper, we propose and discuss a novel approach for measuring perceived contrast. The proposed method comes from the modification of previous algorithms with a different local measure of contrast and with a parameterized way to recombine local contrast maps and color channels. We propose the idea of recombining the local contrast maps using gaze information, saliency maps and a gaze-attentive fixation finding engine as weighting parameters giving attention to regions that observers stare at, finding them important. Our experimental results show that contrast measures cannot be improved using different weighting maps as contrast is an intrinsic factor and its judged by the global impression of the image.

Evaluation of a difference of Gaussians based image difference metric in relation to perceived compression artifacts

In this paper we investigate if the Difference of Gaussians model is able to predict observers perceived difference in relation to compression artifacts. A new image difference metric for specifically designed for compression artifacts is proposed. In order to evaluate this new metric a psychophysical experiment is carried out, where a dataset of 80 compressed JPEG and JPEG2000 images were generated from 10 different scenes. The results of the psychophysical experiment with 18 observers and the quality scores obtained from a large number of image difference metrics are presented. Furthermore, a quantitative study based on a number of image difference metrics and five additional databases is performed in order to reveal the potential of the proposed metric. The analyses show that the proposed metric and most of the tested ones do not correlate well with the subjective test results, and thus the increased complexity of the recent metrics is not justified.

Point-based spatial colour sampling in Milano-Retinex: a survey

Milano-Retinex is a family of Retinex-inspired spatial colour algorithms mainly developed for colour image enhancement. According to the Retinex theory, a Milano-Retinex algorithm takes as input an RGB image and processes the colour intensities of each pixel (i.e. the target) based on the spatial distribution of the colour intensities sampled in a surrounding region. The output is an RGB image, with locally adjusted colours and contrast. In Milano-Retinex family, different ways of spatial sampling are implemented. This study reviews and compares these sampling characteristics within a group of Milano-Retinex algorithms developed in the last decade, from Random Spray Retinex (2007) to the gradient-based colour sampling schemes GREAT and GRASS (2017). Instead of exploring the target neighbourhood by random paths as the original Retinex algorithm does, these methods consider sets of pixels, randomly or deterministically defined, including all the image pixels or a part of them, such as random sprays or image edges. They replace the ratio-reset-threshold-product-average mechanism of the original Retinex with equations re-working maximal intensities over the sampled sets. The performance of these approaches is compared with more than 200 images of indoor and outdoor scenes, captured by commercial cameras under several different conditions.