As large-scale direct view TV screens such as LCD flat panels or plasma displays become more and more affordable, consumers not only expect to buy a ‘big screen’ but to also get ‘great picture quality’. To enjoy the big picture, its viewing distance is significantly reduced. Consequently, more artifacts related to digital compression techniques show above the threshold of visual detectability. The artifact that caught our attention can be noticed within uniform color patches. It presents itself as ‘color blobs’ or color pixel clustering. We analyze the artifact’s color characteristics in RGB and CIELAB color spaces and underline them by re-synthesizing an artificial color patch. To reduce the visibility of the artifact, we elaborate several linear methods, such as low pass filtering and additive white gaussian noise and verify, whether they could correct or mask the visible artifacts. From the huge list of nonlinear filter methods we analyze the effect of high frequency dithering and pixel shuffling, also based on the idea that spatial visual masking should dominate signal correction. By applying shuffling, we generate artificial high frequency components within the uniform color patch. As a result, the artifact characteristics change significantly and its visibility is strongly reduced.
Objective image quality assessment plays an important role in various image processing applications, where the goal of this process is to automatically evaluate the image quality in agreement with human visual perception. In this paper, we propose three different nonlinear learning approaches in order to design image quality assessment models, which serve to predict the perceived image quality. The nonlinear learning approaches used for the aforementioned purpose are nonlinear regression, artificial neural network and regression tree. The largest publicly available image quality database TID2013 is used to benchmark and evaluate the prediction models. The image quality metrics, provided by this TID2013, are not independent and have the redundant information of image quality. This issue might have a negative impact on the training performance and cause overfitting. To avoid this problem and to simplify the model structure, we select the most significant image quality metrics, based on Pearson's correlation measure and principal component analysis. Simulation results confirm that the three nonlinear learning models have high efficiency in predicting image quality. In addition, the regression tree model has low complexity and easy implementation, comparing to the two other prediction models.
Cameras as well as displays of mobile phones, autonomously driven vehicles, PC monitors, and TVs continue to increase their native resolution to 4k by 2k and beyond. At the same time, their high dynamic range formats demand higher bit depth for the underlying color component signals. Subsequently, uncompressed pixel amplitude processing becomes costly not only when transmitting over cable or wireless communication channels, but also across on-chip image processing pipelines that access external memory units. In 2016 we introduced a low cost, real time, visually lossless color image compression concept inspired by structure tensor analysis which promises a highly adaptive and robust compression performance across a substantial range of compression ratios (between 1x and 3x) without significantly compromising perceptual image quality. We also noticed surprisingly strong perceptual color stability in spite of having processed each color component independently in RGB color space. To manage a wider range of compression ratios as well as visually lossless image quality, we proposed a novel approach that converts image amplitudes into a pair of discrete structure and magnitude quantities on a pixel-by-pixel basis which had been inspired by structure tensor analysis. Graceful degradation of image information is controlled by a single parameter which aims at optimally defining sparsity – as a function of image content. Furthermore, we applied error diffusion via a threshold matrix to optimally diffuse the residual coding error. Strongly encouraged by these findings, we continued implementing a version which combines structurally similar elements across RGB color components. As a result, we already achieve visually lossless compression with compression ratios above 4x with 8bit gamma pre-corrected color component signals while having to only analyze 4 nearest neighbors per pixel. We believe to have well identified a conceptual explanation for the algorithm's extraordinary perceptual color stability which we would like to present and discuss in detail. We also provide a detailed error distribution analysis across a variety of well-known, full-reference metrics which highlights the effectiveness of our new approach, identifies its current limitations with regard to high quality color rendering, and illustrates algorithm specific visual artifacts.
This paper proposes two novel approaches to Video Quality Assessment (VQA). Both approaches attempt to develop video evaluation techniques capable of replacing human judgment when rating video quality in subjective experiments. The underlying study consists of selecting fundamental quality metrics based on Human Visual System (HVS) models and using artificial intelligence solutions as well as advanced statistical analysis. This new combination enables suitable video quality ratings while taking as input multiple quality metrics. The first method uses a neural network based machine learning process. The second method consists in evaluating the video quality assessment using non-linear regression model. The efficiency of the proposed methods is demonstrated by comparing their results with those of existing work done on synthetic video artifacts. The results obtained by each method are compared with scores from a database resulting from subjective experiments.
Cet article a pour but de definir les deux principales causes du flou sur les ecrans plats de type LCD et de proposer une solution pour chacun des problemes. Nous indroduirons tout d'abord certaines notions sur les ecrans LCD avant de nous pencher sur le terme de reponse, expression trop souvent employee a mauvaise escient. Nous montrerons ensuite qu'il existe une solution simple pour diminuer ce temps de reponse et nous presenterons enfin la principale cause du flou sur les ecrans LCD de grande taille ainsi qu'une solution adaptee a ce dernier probleme.
During recent years color reproduction systems for consumer needs have experienced various difficulties. In particular, flat panels and printers could not reach a satisfactory color match. The RGB image stored on an Internet server of a retailer did not show the desired colors on a consumer display device or printer device. STMicroelectronics addresses this important color reproduction issue inside their advanced display engines using novel algorithms targeted for low cost consumer flat panels. Using a new and genuine RGB color space transformation, which combines a gamma correction Look-Up-Table, tetrahedrization, and linear interpolation, we satisfy market demands.
A new single chip digital scart interface has been designed to support all the features of digital high end TV receivers. The IC contains the entire circuitry to interface analog YUV/RGB/fast blank signals to a digital YUV system. In particular, it enables inclusion of analog RGB component signals from an external video source in parallel to the primary video signal. The fast blank signal is used to control a soft mixer between the digitized RGB (4:4:4) and an external digital YUV (4:2:2) source. Suitable applications include sample rate conversion, 16:9 display format, and 100 Hz “flicker free” display.
A display’s color subpixel geometry provides an intriguing opportunity for improving readability of text. True type fonts can be positioned at the precision of subpixel resolution. With such a constraint in mind, how does one need to design font characteristics? On the other hand, display manufactures try hard in addressing the color display’s dilemma: smaller pixel pitch and larger display diagonals strongly increase the total number of pixels. Consequently, cost of column and row drivers as well as power consumption increase. Perceptual color subpixel rendering using color component subsampling may save about 1/3 of color subpixels (and reduce power dissipation). This talk will try to elaborate the following questions, based on simulation of several different layouts of subpixel matrices: Up to what level are display device constraints compatible with software specific ideas of rendering text? How much of color contrast will remain? How to best consider preferred viewing distance for readability of text? How much does visual acuity vary at 20/20 vision? Can simplified models of human visual color perception be easily applied to text rendering on displays? How linear is human visual contrast perception around band limit of a display’s spatial resolution? How colorful does the rendered text appear on the screen? How much does viewing angle influence the performance of subpixel layouts and color subpixel rendering?
Large-scale, direct view TV screens, in particular those based on liquid crystal technology, are beginning to use LED (Light Emitting Diode) backlight technology. Conservative estimates show that LED-LCD TVs additional cost will be compensated by reduced power consumption over average operational lifetime. Local dimming not only promises power savings but also improving additional features such as overall image contrast and color gamut. Based on a simplified human visual model with regard to content dependent color discrimination and local adaptation possible tradeoffs with regard to color uniformity, viewing angle, local contrast, and brightness variations are being discussed. Comparing an 'ideal reference' (input to elaborated model) with the output of a tunable model enabled estimating a threshold of artifact visibility for still images as well as video clips that were considered relevant. Overall image quality with regard to dynamic contrast as well as dynamic color gamut, spatially and temporally, was optimized after having obtained visibility thresholds. Finally, some cost functions are proposed that enable optimizing most important color quality parameters.
High-end PC monitors and TVs continue to increase their native display resolution to 4k by 2k and beyond. Subsequently, uncompressed pixel amplitude processing becomes costly not only when transmitting over cable or wireless communication channels, but also when processing with array processor architectures. This paper follows a series of papers we presented earlier on a 4*4 block-based memory compression architecture for text, graphics, and video using a multi-dimensional vector representation with context sensitive control of visually noticeable artifacts. A key feature in the system is the sorting by magnitude of pixel amplitudes. To increase the compression ratio and simultaneously alleviate the limitation on block size, we analyze to which extent the sorting orders can be predicted and we consequently propose new schemes to transmit them efficiently. Depending on the compression ratio, the new cost function defined can be considered as a no-reference or reduced-reference ranking naturalness criterion. We show how pertinent our approach is to additionally correct specific visually noticeable compression artefacts thanks to its adaptive pixel positioning mechanism. Finally, we also provide hints on how to extend this new philosophy to support the optimization of future scalable architectures for transcoding or rendering on high quality displays.
