Hierarchical vectorization for facial images

In this paper, we propose two novel (non- parametric) statistical measures of image and video quality. Both of these are based on the fundamental notion of singular value decomposition, applied to local pixel and higher order information derived from the data. These methods enable measurement of similarity between images, and the measurement of image quality without a reference. A possible derivative application of the latter statistical measure is to optimize parameters of the relevant restoration algorithms being applied.

Clip art databases can be composed by raster images or by vector drawings. There are technologies for searching and retrieving clip arts for both image formats but research has been done separately, focusing on either format, without taking benefits of both research fields as a whole. This paper describes a study where the benefits of combining information extracted from vector and raster images to retrieve clip arts are evaluated and discussed. Color and texture features are extracted from raster images and geometry and topology features are extracted from vector images. The paper presents several comparisons between different combinations with several descriptors. The results of the study show the effectiveness of the solutions that combines both types of features.

The method of object localization of semantic blocks on raster images is considered. The basis of an offered method are results of analysis of adaptive pyramidal representation of raster images. The feature of the algorithm of forming pyramidal representation is the relation between last level of hierarchy value with the level of semantic value that allows to find the areas representing the greatest interest in human perception and to use such selection for dynamic control of a compression.

Color transfer changes the color contents of a target image by replacing the colors in the target image with colors from another source image. The target image will be repainted/recolored to exhibit the same ambience as the source image. Color transfer is applicable to a wide range of commercial image processing tools and products. While much outstanding research has been conducted on this subject, judging the performance of the recoloring process remains subjective to human evaluation. To obtain an objective quantitative assessment of recoloring algorithms' performance, a new color transfer quality measure is proposed. In this paper, we will first establish the requirements that a good color transfer quality measure should meet. Then, according to these requirements, a new color transfer quality measure is proposed that focuses on measuring the content similarity between a target image and a resulting recolored output image, and the color similarity between the source image and the output image. To demonstrate the performance of the proposed measure, the subjective human perception Mean Opinion Score (MOS) values are used. The high correlation between MOS and the proposed measure demonstrate the measure's performance and demonstrate that it exhibits high consistency with human perception.

Combining multiple color image attribute measures by linear addition has been shown to be an effective method in evaluating color image and video quality. Therefore, appropriate selection of the linear combination coefficients determines the performance of the image quality measure. In this paper, training methods of obtaining the linear combination coefficients are discussed for two practical applications: (1) for image processing applications where images before and after processing have the same source; and (2) for video processing purpose where consecutive frames have different contents and suffer from different types of distortions. With the obtained linear combination coefficients, the overall image quality measure Color Quality Measure (CQM) is shown to be effective in benchmarking the qualities of images and video frames in terms that it has high correlations with subjective evaluation Mean Opinion Scores (MOS).

Image Quality Assessment has diverse applications. A number of Image Quality measures are proposed, but none is proved to be true representative of human perception of image quality. We have subjectively investigated spectral distance based and human visual system based image quality measures for their effectiveness in representing the human perception for images corrupted with white noise. Each of the 160 images with various degrees of white noise is subjectively evaluated by 50 human subjects, resulting in 8000 human judgments. On the basis of evaluations, image independent human perception values are calculated. The perception values are plotted against spectral distance based and human visual system based image quality measures. The performance of quality measures is determined by graphical observations and polynomial curve fitting, resulting in best performance by Human Visual System Absolute norm.

Color transfer methods alter the look of a source image with regards to a reference image. So far, the proposed color transfer methods have been limited to low-dynamic-range (LDR) images. Unlike LDR images, which are display-dependent, high-dynamic-range (HDR) images contain real physical values of the world luminance and are able to capture high luminance variations and finest details of real world scenes. Therefore, there exists a strong discrepancy between the two types of images. In this paper, we bridge the gap between the color transfer domain and the HDR imagery by introducing HDR extensions to LDR color transfer methods. We tackle the main issues of applying a color transfer between two HDR images. First, to address the nature of light and color distributions in the context of HDR imagery, we carry out modifications of traditional color spaces. Furthermore, we ensure high precision in the quantization of the dynamic range for histogram computations. As image clustering (based on light and colors) proved to be an important aspect of color transfer, we analyze it and adapt it to the HDR domain. Our framework has been applied to several state-of-the-art color transfer methods. Qualitative experiments have shown that results obtained with the proposed adaptation approach exhibit less artifacts and are visually more pleasing than results obtained when straightforwardly applying existing color transfer methods to HDR images. © (2015) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.

In this paper, we present a generative sketch model for human hair analysis and synthesis. We treat hair images as2D piecewise smooth vector (flow) fields and, thus, our representation is view-based in contrast to the physically-based 3D hair models in graphics. The generative model has three levels. The bottom level is the high-frequency band of the hair image. The middle level is a piecewise smooth vector field for the hair orientation, gradient strength, and growth directions. The top level is an attribute sketch graph for representing the discontinuities in the vector field.Asketch graph typically has a number of sketch curves which are divided into 11 types of directed primitives. Each primitive is a small window (say 5 _ 7 pixels) where the orientations and growth directions are defined in parametric forms, for example, hair boundaries, occluding lines between hair strands, dividing lines on top of the hair, etc. In addition to the three level representation, we model the shading effects, i.e., the low-frequency band of the hair image, by a linear superposition of some Gaussian image bases andwe encode the hair color by a color map. The inference algorithm is divided into two stages: 1) We compute the undirected orientation field and sketch graph from an input image and 2) we compute the hair growth direction for the sketch curves and the orientation field using a Swendsen-Wang cut algorithm. Both steps maximize a joint Bayesian posterior probability. The generative model provides a straightforward way for synthesizing realistic hair images and stylistic drawings (rendering) from a sketch graph and a few Gaussian bases. The latter can be either inferred from a real hair image or input(edited) manually using a simple sketching interface. We test our algorithm on a large data set of hair images with diverse hair styles.

To provide semantic image style transfer results which are consistent with human perception, transferring styles of semantic regions of the style image to their corresponding semantic regions of the content image is necessary. However, when the object categories between the content and style images are not the same, it is difficult to match semantic regions between two images for semantic image style transfer. To solve the semantic matching problem and guide the semantic image style transfer based on matched regions, we propose a novel semantic context-aware image style transfer method by performing semantic context matching followed by a hierarchical local-to-global network architecture. The semantic context matching aims to obtain the corresponding regions between the content and style images by using context correlations of different object categories. Based on the matching results, we retrieve semantic context pairs where each pair is composed of two semantically matched regions from the content and style images. To achieve semantic context-aware style transfer, a hierarchical local-to-global network architecture, which contains two sub-networks including the local context network and the global context network, is proposed. The former focuses on style transfer for each semantic context pair from the style image to the content image, and generates a local style transfer image storing the detailed style feature representations for corresponding semantic regions. The latter aims to derive the stylized image by considering the content, the style, and the intermediate local style transfer images, so that inconsistency between different corresponding semantic regions can be addressed and solved. The experimental results show that the stylized results using our method are more consistent with human perception compared with the state-of-the-art methods.

EDITORIAL article Front. Pediatr., 05 November 2015Sec. Neonatology Volume 3 - 2015 | https://doi.org/10.3389/fped.2015.00095

The color logarithmic image processing (CoLIP) is a mathematical framework for the representation and processing of color images. It is psychophysically well justified since it is consistent with several human visual perception laws and characteristics. It is mathematically and computationally relevant since it allows to consider color images as vectors in an abstract linear space, contrary to the classical color spaces (e.g., RGB and \(L^*a^*b^*\)). The first purpose of this chapter is to present the mathematical fundamentals of the CoLIP together with its main psychophysical connections (Grasmann’s law, color matching functions, chromaticity diagram, and the Maxwell triangle). The second purpose is to present some basic image processing and analysis techniques for contrast enhancement (histogram equalization, dynamic range maximization, and toggle contrast calculation), white balance correction, color transfer, K-means clustering, and filtering. Most of them are applied on various original color images in a comparative way between CoLIP, RGB, and \(L^*a^*b^*\) color spaces.

Underwater images suffer from blurring effects, low contrast, and grayed out colors due to the absorption and scattering effects under the water. Many image enhancement algorithms for improving the visual quality of underwater images have been developed. Unfortunately, no well-accepted objective measure exists that can evaluate the quality of underwater images similar to human perception. Predominant underwater image processing algorithms use either a subjective evaluation, which is time consuming and biased, or a generic image quality measure, which fails to consider the properties of underwater images. To address this problem, a new nonreference underwater image quality measure (UIQM) is presented in this paper. The UIQM comprises three underwater image attribute measures: the underwater image colorfulness measure (UICM), the underwater image sharpness measure (UISM), and the underwater image contrast measure (UIConM). Each attribute is selected for evaluating one aspect of the underwater image degradation, and each presented attribute measure is inspired by the properties of human visual systems (HVSs). The experimental results demonstrate that the measures effectively evaluate the underwater image quality in accordance with the human perceptions. These measures are also used on the AirAsia 8501 wreckage images to show their importance in practical applications.

A crucial deficiency of lossy blockwise image compression is the generation of local artifacts such as ringing defects, obscuration of fine detail, and blocking effect (BE). To date, few published reports of image quality measures (IQMs) have addressed the detection of such errors in a realistic, efficient manner. Exceptions are feature-based IQMs, perceptual IQMs, error detection templates, and quantification of BE that support its reduction in JPEG-and wavelet-compressed imagery. In this paper, we present an enhanced suite of IQMs that emphasize detection of local, feature-specific errors that corrupt visual appearance or numerical integrity of decompressed digital imagery. By the term visual appearance is meant subjective error, in contrast with objectively quantified effects of compression on individual pixel values and their spatial interrelationships. Subjective error is of key importance in human viewing applications, for examples, Internet video. Objective error is primarily of interest in object recognition applications such as automated target recognition (ATR), where implementational concerns involve the effect of compression or decompression algorithms on probability of detection (Pd) and rate of false alarms (Rfa). Analysis of results presented herein emphasizes application- specific quantification of local compression errors. In particular, introduction of extraneous detail (e.g., ringing defects of BE) or obscuration of source detail (e.g., texture masking) adversely impact both subjective and objective error of a decompressed image. Blocking effect is primarily a visual problem, but can confound ATR filters when a target spans a block boundary. Introduction of point or cluster errors primarily degrades ATR filter performance, but can also produce noticeable degradation of fine detail for human visual evaluation of decompressed imagery. In practice, error and performance analysis is supported by examples of ATR imagery including airborne and underwater mono-and multi-spectral images.

Proper use of functional neuro-imaging through effective experimental design and modern statistical analysis provides new insights in current brain research. This tutorial has two aims: to describe aspects of this technology to applied statisticians and to provide some statistical ideas to neuroscientists unfamiliar with quantitative analytic methods that accommodate randomness. Introductory background material and ample references to current literature on the physics of magnetic resonance imaging, Fourier methods for image reconstruction and measures of image quality are included. Two of the statistical approaches mentioned here are extensions of established methods for longitudinal data analysis to the frequency domain. A recent case study provides real-world instances of approaches, problems and open questions encountered in current functional neuro-imaging research and an introduction to the analysis of spatial time series in this context.

Image and video capture devices are often designed with different capabilities and end-users in mind. As a result, very often images are captured at one resolution and need to be displayed at another resolution. Resizing thus continues to be an important research area in today's milieu. Well-known pixel-based methods (like bicubic interpolation) have traditionally been used for the purpose. However, these methods are often found lacking in quality from a perceptual standpoint as they tend to soften edges and blur details. Recently, vectorization based approaches have been explored because of their inherent property of artifact free scaling. However, these methods fail to faithfully reproduce textures and fine details in an image. In this work, we present a novel, layered approach of combining image vectorization with a pixel based interpolation technique to achieve high quality scaling of digital images. The proposed method decomposes an image into two layers: a `coarse' layer capturing the visually important structure of the image, and a `fine' layer containing the details. We vectorize only the coarse layer and render at the desired scale, while using classical interpolation on the fine layer. The final scaled image is composed by blending the independently scaled layers. We compare the performance of the proposed method with several state-of-the-art vectorization and scaling methods, and report comparably better performance.