Random Spray Retinex Research Articles

An Efficient In-Memory Computing Architecture for Image Enhancement in AI Applications

Random Spray Retinex (RSR) is an effective image enhancement algorithm due to its effectiveness in improving image quality. However, the algorithm computing complexity and the required hardware resources and memory accesses hampered its deployment in many applications scenarios, for instance in IoT systems with limited hardware resources. With the rise of Artificial Intelligence (AI), the use of image enhancement has become essential to improve the performance for many emerging applications. In this paper, we propose the use of the RSR as a pre-processing filter before the task of semantic segmentation of low-quality urban road scenes. Using the publicly available Cityscapes dataset, we compare the performance of a pre-trained deep semantic segmentation network on dark noisy images and on RSR pre-processed images. Our findings confirm the effectiveness of RSR in improving segmentation accuracy. In addition, to address the computation complexity and suitability to edge devices, we propose a novel efficient implementation of the RSR using resistive random access memory (RRAM) technology. The architecture provides highly parallel analog in-memory computing (IMC) capabilities. A detailed, efficient and low latency implementation of the RSR using RRAM-CMOS technology is described. The design is verified using SPICE simulations with measured data from fabricated RRAM and 65nm CMOS technologies. The approach provided here represents an important step towards low-complexity and real-time hardware-friendly architecture and design for Retinex algorithms for edge devices.

Developing an inkjet printer IV: printer mechanism control for best print quality

The Random spray Retinex (RSR) algorithm was developed by taking into consideration the mathematical description of Milano-Retinex. The RSR substituted random paths with random sprays. Mimicking some characteristics of the human visual system (HVS), this article proposes two variants of RSR adding a mechanism of region of interest (ROI). In the first proposed model, a cone distribution based on anatomical data is considered as ROI. In the second model, the visual resolution depending on the visual field based on the knowledge of visual information processing is considered as ROI. We have measured actual eye movements using an eye-tracking system. By using the eye-tracking data, we have simulated the HVS using test images. Results show an interesting qualitative computation of the appearance of the processed area around real gaze points.

Random Spray Retinex Extensions Considering Region of Interest and Eye Movements

Abstract The Random spray Retinex (RSR) algorithm was developed by taking into consideration the mathematical description of Milano-Retinex. The RSR substituted random paths with random sprays. Mimicking some characteristics of the human visual system (HVS), this article proposes two variants of RSR adding a mechanism of region of interest (ROI). In the first proposed model, a cone distribution based on anatomical data is considered as ROI. In the second model, the visual resolution depending on the visual field based on the knowledge of visual information processing is considered as ROI. We have measured actual eye movements using an eye-tracking system. By using the eye-tracking data, we have simulated the HVS using test images. Results show an interesting qualitative computation of the appearance of the processed area around real gaze points.

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.

A fuzzy set approach to Retinex spray sampling

The color sensation at a point, for the Human Visual System (HVS), derives not only from the color stimulus at that point, but also from the relative spatial arrangement of the stimuli in the image. Based on this observation, the Retinex algorithm, an early and widely studied model of the HVS, determines the output – for each chromatic channel – by rescaling the input intensity of a pixel w.r.t. a reference white level, computed by sampling the brightest points in the neighborhood of the target pixel. In this work, we argue that several elements, inherent to the above observation, can benefit from a fuzzy formalization. We show that the adoption of the fuzzy formalism allows to better encode the mutual influence of pixels. Overall, the fuzzy formalization can provide a general framework for designing and tuning image enhancement algorithms inspired by the HVS. We demonstrate its use by the construction of a fuzzy version of the point-sampling algorithm Random Spray Retinex (RSR). Using RSR as a guide, we build a more efficient algorithm, based on the fact that each spray (a set of sampled points used in RSR to determine the reference white of a specific target) can be assumed to belong to some degree to all the target pixels of the image, provided that a suitable membership function is defined. The features of this alternative formalization of RSR are discussed here, using synthetic and natural test images.

From samples to populations in retinex models

Some spatial color algorithms, such as Brownian Milano retinex (MI-retinex) and random spray retinex (RSR), are based on sampling. In Brownian MI-retinex, memoryless random walks (MRWs) explore the neighborhood of a pixel and are then used to compute its output. Considering the relative redundancy and inefficiency of MRW exploration, the algorithm RSR replaced the walks by samples of points (the sprays). Recent works point to the fact that a mapping from the sampling formulation to the probabilistic formulation of the corresponding sampling process can offer useful insights into the models, at the same time featuring intrinsically noise-free outputs. The paper continues the development of this concept and shows that the population-based versions of RSR and Brownian MI-retinex can be used to obtain analytical expressions for the outputs of some test images. The comparison of the two analytic expressions from RSR and from Brownian MI-retinex demonstrates not only that the two outputs are, in general, different but also that they depend in a qualitatively different way upon the features of the image.

Tuning the locality of filtering with a spatially weighted implementation of random spray Retinex

The human color sensation depends on the spatial distribution of the colors in the viewed scene. This principle is at the basis of the random spray Retinex (RSR) algorithm. In this work, we modify RSR by integrating its approach with a method to weight and tune the locality of spatial image information. This modification allows for spatial control of the local effect of RSR on image color filtering. We study the performances of this spatially weighted version of RSR on a public image dataset by analyzing and comparing several image features of the output image and its local properties.

QBRIX: a quantile-based approach to retinex.

In this paper, we introduce a novel probabilistic version of retinex. It is based on a probabilistic formalization of the random spray retinex sampling and contributes to the investigation of the spatial properties of the model. Various versions available of the retinex algorithm are characterized by different procedures for exploring the image content (so as to obtain, for each pixel, a reference white value), then used to rescale the pixel lightness. Here we propose an alternative procedure, which computes the reference white value from the percentile values of the pixel population. We formalize two versions of the algorithm: one with global and one with local behavior, characterized by different computational costs.

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 (Random Spray Retinex (RSR) )and an essentially GW one (Automatic Color Equalization (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.

Random Spray Retinex: A New Retinex Implementation to Investigate the Local Properties of the Model

In order to investigate the local filtering behavior of the Retinex model, we propose a new implementation in which paths are replaced by 2-D pixel sprays, hence the name "random spray Retinex." A peculiar feature of this implementation is the way its parameters can be controlled to perform spatial investigation. The parameters' tuning is accomplished by an unsupervised method based on quantitative measures. This procedure has been validated via user panel tests. Furthermore, the spray approach has faster performances than the path-wise one. Tests and results are presented and discussed.