Alessandro Rizzi

Mathematical definition and analysis of the Retinex algorithm

We present a detailed mathematical analysis of the original Retinex algorithm due to Land and McCann [J. Opt. Soc. Am. 61, 1 (1071)]. To this end, we propose an analytic formula that describes the algorithm behavior. More than one Retinex version (e.g., with and without threshold) is examined. The behavior of Retinex varying the number of paths is predicted, and its recursive iterations are mathematically analyzed using the formula. The mathematical setting presented serves as a common ground for the various Retinex implementations. Its validity is confirmed by the tests on images that we have performed.

A Spatially Variant White-Patch and Gray-World Method for Color Image Enhancement Driven by Local Contrast

Starting from the revolutionary Retinex by Land and McCann, several further perceptually inspired color correction models have been developed with different aims, e.g. reproduction of color sensation, robust features recognition, enhancement of color images. Such models have a differential, spatially-variant and non-linear nature and they can coarsely be distinguished between white-patch (WP) and gray-world (GW) algorithms. In this paper we show that the combination of a pure WP algorithm (RSR: random spray Retinex) and an essentially GW one (ACE) leads to a more robust and better performing model (RACE). The choice of RSR and ACE follows from the recent identification of a unified spatially-variant approach for both algorithms. Mathematically, the originally distinct non-linear and differential mechanisms of RSR and ACE have been fused using the spray technique and local average operations. The investigation of RACE allowed us to put in evidence a common drawback of differential models: corruption of uniform image areas. To overcome this intrinsic defect, we devised a local and global contrast-based and image-driven regulation mechanism that has a general applicability to perceptually inspired color correction algorithms. Tests, comparisons and discussions are presented.

From Retinex to Automatic Color Equalization: issues in developing a new algorithm for unsupervised color equalization

We present a comparison between two color equalization algorithms: Retinex, the famous model due to Land and McCann, and Automatic Color Equalization (ACE), a new algorithm recently presented by the authors. These two algorithms share a common approach to color equalization, but different computational models. We introduce the two models focusing on differences and common points. An analysis of their computational characteristics illustrates the way the Retinex approach has influenced ACE structure, and which aspects of the first algorithm have been modified in the second one and how. Their interesting equalization properties, like lightness and color constancy, image dynamic stretching, global and local filtering, and data driven dequantization, are qualitatively and quantitatively presented and compared, together with their ability to mimic the human visual system.

Perceptual Color Correction Through Variational Techniques

In this paper, we present a discussion about perceptual-based color correction of digital images in the framework of variational techniques. We propose a novel image functional whose minimization produces a perceptually inspired color enhanced version of the original. The variational formulation permits a more flexible local control of contrast adjustment and attachment to data. We show that a numerical implementation of the gradient descent technique applied to this energy functional coincides with the equation of automatic color enhancement (ACE), a particular perceptual-based model of color enhancement. Moreover, we prove that a numerical approximation of the Euler-Lagrange equation reduces the computational complexity of ACE from O(N2) to O(NlogN), where N is the total number of pixels in the image

A computational approach to color adaptation effects

Abstract The human vision system has adaptation mechanisms that cannot be managed with the classic tri-stimulus color theory. The effects of these mechanisms are clearly visible in some well-known perception phenomena as color illusions, but they are always present in human observation. The discrepancy between the observation of a real scene and the observation of a picture taken from the same scene, derives from the fact that the camera does not have such mechanisms. In this paper, we propose a biologically inspired implementation of the Retinex algorithm, introduced by Land and McCann, that simulates these adaptation mechanisms, in order to reproduce some effects like dynamic adjustment, color constancy, etc. typical of the human vision system. The algorithm has been tested not only on a Mondrian-like patchwork to measure its effect, but also on different pictures, photographs and typical color illusions to test its adaptation effects. The examples demonstrate the ability of the model to emulate some characteristics of human color perception and to obtain better equalized and color-corrected images.

A Multiscale Framework for Spatial Gamut Mapping

Image reproduction devices, such as displays or printers, can reproduce only a limited set of colors, denoted the color gamut. The gamut depends on both theoretical and technical limitations. Reproduction device gamuts are significantly different from acquisition device gamuts. These facts raise the problem of reproducing similar color images across different devices. This is well known as the gamut mapping problem. Gamut mapping algorithms have been developed mainly using colorimetric pixel-wise principles, without considering the spatial properties of the image. The recently proposed multilevel gamut mapping approach takes spatial properties into account and has been demonstrated to outperform spatially invariant approaches. However, they have some important drawbacks. To analyze these drawbacks, we build a common framework that encompasses at least two important previous multilevel gamut mapping algorithms. Then, when the causes of the drawbacks are understood, we solve the typical problem of possible hue shifts. Next, we design appropriate operators and functions to strongly reduce both haloing and possible undesired over compression. We use challenging synthetic images, as well as real photographs, to practically show that the improvements give the expected results.

Camera and visual veiling glare in HDR images

— High-dynamic-range (HDR) images are superior to conventional images. The experiments in this paper measure camera and human responses to calibrated HDR test targets. We calibrated a 4.3-log-unit test target, with minimal and maximal glare from a changeable surround. Glare is an uncontrolled spread of an image-dependent fraction of scene luminance in cameras and in the eye. We use this standard test target to measure the range of luminances that can be captured on a cameras image plane. Further, we measure the appearance of these test luminance patches. We discuss why HDR is better than conventional imaging, despite the fact the reproduction of luminance is inaccurate.

Analgesia in patients undergoing thoracotomy: Epidural versus paravertebral technique. A randomized, double-blind, prospective study

BACKGROUND Pain control after thoracotomy prevents postsurgical complications and improves respiratory function. The gold standard for post-thoracotomy analgesia is the epidural catheter. The aim of this study was to compare it with a new technique that involves placement of a catheter in the paravertebral space at the end of surgery under a surgeons direct vision. METHODS From November 2011 to June 2012, 52 patients were randomized into 2 groups depending on catheter placement: an epidural catheter for group A and a paravertebral catheter for group B. At 12, 24, 48, and 72 hours after surgery, the following parameters were recorded: (1) pain control using the patients completion of a visual analog scale module, (2) respiratory function using forced expiratory volume in 1 second and ambient air saturation, and (3) blood cortisol values as an index of systemic reaction to pain. RESULTS Statistically significant differences (P < .05) were found in favor of group B for both cough and rest pain control (P = .002 and .002, respectively) and respiratory function in terms of forced expiratory volume in 1 second and ambient air saturation levels (P = .023 and .001, respectively). No statistically significant differences were found in blood cortisol trends between the 2 groups (P > .05). Collateral effects such as vomiting, nausea, low pressure, or urinary retention were observed only in group A. No collateral effects were recorded in the paravertebral group. CONCLUSIONS According to our data, drugs administered through a paravertebral catheter are very effective. Moreover, it does not present contraindications to its positioning or collateral effects. More studies are necessary to confirm data we collected.

3D modeling of complex and detailed cultural heritage using multi-resolution data

The article reports the interdisciplinary project of the virtualization of the Great Inscription of Gortyna, Crete, for 3D documentation, structural studies, and physical replica purposes. The digitization of the longest epigraphic text of the Greek civilization (6 m long and 1.75 m high, with approximately 2--3 mm-depth engraved letters) and its surrounding heritage area (around 30 × 30 m), required long planning and the construction of a dedicated acquisition system to speed up the surveying time, limited to few hours per day. Primarily, range sensors were employed in a multi-resolution way, digitizing detailed parts in high resolution and less smoothed areas with lower geometric resolution. Some selected areas were also modeled with our multiphoto geometrically constrained image matching approach to demonstrate that the same accuracy and details can be achieved using either scanners or photogrammetry. The derived 3D model of the heritage is now the basis for further archaeological studies on the incision techniques and a deeper structural analysis on the monument. The challenges of the work stay in the acquisition, processing, and integration of the multi-resolution data as well as their interactive visualization.

Color correction between gray world and white patch

Color equalization algorithms exhibit a variety of behaviors described in two differing types of models: Gray World and White Patch. These two models are considered alternatives to each other in methods of color correction. They are the basis for two human visual adaptation mechanisms: Lightness Constancy and Color Constancy. The Gray World approach is typical of the Lightness Constancy adaptation because it centers the histogram dynamic, working the same way as the exposure control on a camera. Alternatively, the White Patch approach is typical of the Color Constancy adaptation, searching for the lightest patch to use as a white reference similar to how the human visual system does. The Retinex algorithm basically belongs to the White Patch family due to its reset mechanism. Searching for a way to merge these two approaches, we have developed a new chromatic correction algorithm, called Automatic Color Equalization (ACE), which is able to perform Color Constancy even if based on Gray World approach. It maintains the main Retinex idea that the color sensation derives from the comparison of the spectral lightness values across the image. We tested different performance measures on ACE, Retinex and other equalization algorithms. The results of this comparison are presented.