Color Representation Research Articles

Correlation Properties of Color Histograms in the Case of Image Quality Decreasing

In image recognition tasks, objects are identified by examining various features such as texture, color, contour detection, and statistical or semantic descriptions. One widely used approach for extracting image attributes is the analysis of intensity histograms. While the traditional RGB color model is commonly used in digital image processing, it is often more effective to analyze color properties in HS* systems (such as HSL, HSV, and HSI) since these systems more closely resemble the spectral representation of color. A key characteristic shared by these three systems is the use of the H (Hue) coordinate, which is represented as an angular value within a cylindrical coordinate system. The paper investigates the possibility of using color histograms generated in HS* spaces for identifying images that have undergone various types of distortions. The CQ100: A High-Quality Image Dataset for Color Quantiza-tion Research was chosen for the research. The non-quantized section of the CQ100 dataset consists of 100 RGB images in PNG format, each with a resolution of 768×512 pixels and a color depth of 24 bits. The study examines how different distortions, which can occur during real-time photo and video capture, affect the color properties of images. Specifically, the re-search focuses on distortions caused by rotation, noise, blurring, and optical aberration. His-tograms were compared using the Pearson cross-correlation coefficient, and the findings re-veal that the correlation remains high for the same image despite the applied distortions. Conversely, the correlation coefficient between different images is low for most of the studied objects. These results suggest that color histograms could be effectively used for image identi-fication tasks, even when images are significantly distorted, as is common in image registra-tion processes. The applicability of correlation detection as a method for histograms com-parison is considered regardless of the relative simplicity of its calculation. This approach could contribute to the development of faster image recognition systems.

CSPDAE: Colorization via SuperPixel Downsampler Denoising AutoEncoder towards enhanced color structural clarity

Based on the various structural textures, sizes, and shapes of the gray image, the SuperPixel (SP)-based methods convert the rigid pixels into adaptable image patches that share common features. As a result, the SP-based colorization technique not only enhances the color appearance but also preserves the topological integrity of the images. Despite the effectiveness of neural network-based colorization methods, their integration with SP segmentation has traditionally been complex and cumbersome. To address this issue, we propose the Colorization SP Downsampler Denoising AutoEncoder (CSPDAE), where the SP downsampler integrates SP segmentation directly into the network, eliminating the need for any prior input. The SP downsampler addresses the challenges of computational complexity and small region clustering, which are the main obstacles preventing Transformer architectures from being applied to pixel-level segmentation, through the use of SP cross-attention and aggregated positional embedding (APE). Furthermore, we have incorporated a Color Weight (CW) loss, based on the CIEDE2000 color difference formula, to ensure balanced pixel sampling and to improve the precision of detailed color representation. The experimental results confirm the effectiveness of our method, demonstrating its capacity to produce colors with greater structural accuracy and visually appealing details.

Semantic segmentation using synthetic images of underwater marine-growth.

Subsea applications recently received increasing attention due to the global expansion of offshore energy, seabed infrastructure, and maritime activities; complex inspection, maintenance, and repair tasks in this domain are regularly solved with pilot-controlled, tethered remote-operated vehicles to reduce the use of human divers. However, collecting and precisely labeling submerged data is challenging due to uncontrollable and harsh environmental factors. As an alternative, synthetic environments offer cost-effective, controlled alternatives to real-world operations, with access to detailed ground-truth data. This study investigates the potential of synthetic underwater environments to offer cost-effective, controlled alternatives to real-world operations, by rendering detailed labeled datasets and their application to machine-learning. Two synthetic datasets with over 1000 rendered images each were used to train DeepLabV3+ neural networks with an Xception backbone. The dataset includes environmental classes like seawater and seafloor, offshore structures components, ship hulls, and several marine growth classes. The machine-learning models were trained using transfer learning and data augmentation techniques. Testing showed high accuracy in segmenting synthetic images. In contrast, testing on real-world imagery yielded promising results for two out of three of the studied cases, though challenges in distinguishing some classes persist. This study demonstrates the efficiency of synthetic environments for training subsea machine learning models but also highlights some important limitations in certain cases. Improvements can be pursued by introducing layered species into synthetic environments and improving real-world optical information quality-better color representation, reduced compression artifacts, and minimized motion blur-are key focus areas. Future work involves more extensive validation with expert-labeled datasets to validate and enhance real-world application accuracy.

Color Representation Method of Art Painting Based on Multi-Scale Retinex

Color Representation Method of Art Painting Based on Multi-Scale Retinex

An Intelligent Approach of Underwater Image Co-Enhancement Using Correlation Feature Matching with Enhanced Meta-Heuristic Optimization-Aided Transformer UNet

In marine ecosystems, the geological formation, and aquatic life are analyzed by underwater imaging. The visibility of the aquatic environment is enhanced by the underwater image improvement approaches. Because of the dispersal of the particles, attenuation of light, and distortion of color, the capturing of marine images is a challenging task. The enhanced images with the absence of water turbidity and light attenuation provide an excellent view of the underwater sceneries. The dispersion and incorporation of light underwater cause color distortion effects. Without disturbing the accuracy of the underwater image, the process of obtaining an accurate color representation becomes a complicated task. Overall excellence of the enhanced images may affected by the noise or the artifacts by traditional approaches which tends to affect the image quality. Also, the preservation of important details in the image is a rigorous task while removing the distortion. To combat such issues, an adaptive deep learning-based underwater image co-enhancement is suggested. In the first phase, the images are gathered from the standard dataset. Further, the image is fed into the correlation feature matching module. Here, the feature matching is executed via the Adaptive Trans-Unet (ATUNet), and the correlation feature matching performance of the ATUNet is strengthened by tuning the parameter using the Updated Random Parameter-based Red Kite Optimization Algorithm (URP-RKOA). The usefulness of the developed model is defined using the experimental study. Throughout the analysis, the numerical findings of the developed model attain the values of 155.01, 26.2, and 0.78 in terms of MSE, PSNR, and SSIM. The underwater image co-enhancement achieved better performance rather than existing approaches.

Perceptual categorization and perceptual concepts

Abstract Conceptualism is the view that at least some perceptual representation is conceptual. This paper considers a prominent recent argument against Conceptualism due to Ned Block. Block's argument appeals to patterns of color representation in infants, alleging that infants exhibit categorical perception of color while failing to deploy concepts of color categories. Accordingly, the perceptual representation of color categories in infancy must be non-conceptual. This argument is distinctive insofar as it threatens not only the view that all perception is conceptual, but also views that restrict the Conceptualist thesis to perceptual categorization. However, I contend that it fails at two stages. Block's arguments for the perceptual representation of color categories in infancy, and against color concept deployment in infancy, are unpersuasive. Thus, Block has not vanquished Conceptualism. I draw out implications for debates about the perception–cognition border and for the question of whether explicit categorization occurs in perception.

Using co-design to understand consumer’s health information-seeking behaviours and design preferences for a new digital clinical dashboard in aged care

BackgroundLittle is known about the information needs of older people and their family caregivers, particularly around medication management. This is largely because this population are infrequently consulted in research. Health technologies such as digital dashboards can present comprehensive and timely data summaries to improve knowledge and guide decision-making.ObjectivesThe aim of this study was to understand current information seeking-behaviours of aged care service users and their families, and their preferences for the presentation of these types of information to support the co-design of a digital aged care dashboard.MethodsAged care clients (community and residential aged care) and their family members were invited to participate in semi-structured, one-on-one interviews conducted over the telephone. The interviews explored how participants sought health information, how this information supported their clinical decision-making, and how they would want this information to be presented via a digital dashboard. Interviews were audio-recorded, transcribed verbatim and independently analysed using an inductive content approach by three reviewers to identify prominent categories.ResultsAged care clients and family members sought health information from healthcare providers, aged care facilities and the Internet. Information regarding medications, medical conditions and an overview for family members were of particular importance to participants. Participants saw the usefulness of a digital dashboard displaying relevant information, and provided suggestions regarding presentation of this information. This included the use of large text, representative colours, symbols and graphs. Participants recommended being able to tailor the dashboard to individual end-users.ConclusionsBy engaging aged care clients and their families in research, it was found that they had positive views about a digital dashboard that presented clinical information and was readily accessible. Such a dashboard could complement their current practices of information-seeking. However, for the dashboard to be effective for this population, several suggestions for its design were raised. Understanding the information-seeking behaviours of aged care clients and their families, together with knowledge of the information that is important to them, and then subsequent incorporation of this information into a digital dashboard that reflects their design preferences, could lead to more informed decision-making in this population.

The influence of anxiety on the structure of color associations

The subject of the study is the relationship between the level of anxiety and color associations with anthropologically significant concepts. The goal is an experimental test of the hypothesis that anxiety affects the structure, density and components of the color association system. The experiment involved 100 people (23 men and 77 women) aged 17 to 24 years (mean age 19.71, SD = 1.72). Their anxiety level was established using the Taylor Manifest Anxiety Scale (TMAS) as adapted by V. G. Norakidze. Depending on the total score on the Taylor Manifest Anxiety Scale, we formed 3 groups for participants with low, medium and high anxiety levels. The palette of the experiment included 9 color samples of the NCS natural color system, which were divided into "warm" and "cold". The participants compared the color samples of the palette with 26 anthropologically significant concepts. First, the experiment showed that the anxiety level related to the associative group of preferred colors: as the anxiety level increased, the number of "warm" colors significantly decreased, and "cold" ones increased. Second, in all three groups, "red" and "gray" associations dominated, the differences in the frequency of other selected hues, with the exception of red, blue-green and yellow-green, were quite weak. Third, a correlation between salient individual chromatic characteristics in the color association structure and the anxiety level was revealed: in groups with a higher anxiety level, we found a reduction in the number of connections and an increase in the dominance of individual associations. This study clarifies the relationship between the anxiety level and color cognition, primarily it concerns the color representation of a person’s own state and mood. The research can be continued through the study of other chromatic characteristics, cross-cultural and sociocultural specifics.

Skin Color Representation in Infectious Disease Textbooks.

Darker skin is underrepresented in medical textbooks. This may contribute to inequity in our system. We assessed skin color representation in infectious disease textbooks and compared diseases associated with social stigma to those not stigmatized. Images depicting conditions associated with social stigma were significantly more common on darker skin.

An Analysis of the Diversity of Skin Colour Representation in Paediatric Nursing Practitioner Textbooks.

Our study aims to analyse 8 commonly used textbooks to determine how diverse skin tones are represented in paediatric nursing practitioner education. Literature reviewed from 2016 to 2024 demonstrated that the lack of darkly pigmented skin colour representation in health science education leads to diminished patient outcomes for these populations. Our study sought to study representation teaching images and eight commonly referenced nursing textbooks were chosen for this study, given their use in paediatric nurse practitioner education. Of the eight textbooks selected, five were analysed based on inclusion criteria. Two investigators trained in skin prototyping coded each textbook for skin colour representation and coded during 2023-2024. Coders used the widely accepted prototyping scale, the Fitzpatrick Scale (range I-VI, with I being the lightest colour skin and VI the darkest). Teaching photographs were defined as all photos used to provide insight into a disease or diagnostic technique that included human skin. Two individual coders coded and documented data, ensuring each coder was blinded to the overall results. Our analysis of 5 textbooks revealed that 2112 images met the criteria as teaching images. Of the 2112 teaching images, 593.5 included images of type IV-VI skin (darkly pigmented skin), resulting in a 28% representation of dark skin tone images. Additionally, 2 of the 82 total illustrations included patients with dark skin tones, indicating a representation of 2.5%. However, chapters addressing conditions of child abuse/neglect (55.95%) and stigmatised social issues (infectious disease, 54.88%) displayed a disproportionate representation of patients from these demographics. Our results highlight the importance of enhancing equitable representation in educational resources for nursing practitioners. There is room to collaborate with other health science institutions to establish clear guidelines for future improvement in expanding teaching images to include diversity representation in education. No patient or public contribution.

Advanced Image Preprocessing and Integrated Modeling for UAV Plant Image Classification

The automatic identification of plant species using unmanned aerial vehicles (UAVs) is a valuable tool for ecological research. However, challenges such as reduced spatial resolution due to high-altitude operations, image degradation from camera optics and sensor limitations, and information loss caused by terrain shadows hinder the accurate classification of plant species from UAV imagery. This study addresses these issues by proposing a novel image preprocessing pipeline and evaluating its impact on model performance. Our approach improves image quality through a multi-step pipeline that includes Enhanced Super-Resolution Generative Adversarial Networks (ESRGAN) for resolution enhancement, Contrast-Limited Adaptive Histogram Equalization (CLAHE) for contrast improvement, and white balance adjustments for accurate color representation. These preprocessing steps ensure high-quality input data, leading to better model performance. For feature extraction and classification, we employ a pre-trained VGG-16 deep convolutional neural network, followed by machine learning classifiers, including Support Vector Machine (SVM), random forest (RF), and Extreme Gradient Boosting (XGBoost). This hybrid approach, combining deep learning for feature extraction with machine learning for classification, not only enhances classification accuracy but also reduces computational resource requirements compared to relying solely on deep learning models. Notably, the VGG-16 + SVM model achieved an outstanding accuracy of 97.88% on a dataset preprocessed with ESRGAN and white balance adjustments, with a precision of 97.9%, a recall of 97.8%, and an F1 score of 0.978. Through a comprehensive comparative study, we demonstrate that the proposed framework, utilizing VGG-16 for feature extraction, SVM for classification, and preprocessed images with ESRGAN and white balance adjustments, achieves superior performance in plant species identification from UAV imagery.

Analysis of DD* and D¯(*)Ξcc(*) molecule by one boson exchange model based on heavy quark symmetry

Numerous exotic hadrons with heavy quarks have been reported in the experiments. In such states, symmetries of heavy quarks are considered to play a significant role. In particular, the superflavor symmetry, or also called the heavy quark antidiquark symmetry is one of the interesting symmetries, which links a quark Q to an antidiquark Q¯Q¯ having the same color representation as Q. In this paper, we study a D¯Ξcc molecular state as a superflavor partner of the doubly charmed tetraquark Tcc reported by LHCb recently. Tcc locating slightly below the DD* threshold is a candidate of the hadronic molecule. Thus by replacing the singly charmed meson D(*) with the doubly charmed baryon Ξcc(*), superflavor symmetry predicts the existence of the D¯(*)Ξcc(*) bound state. We employ the one boson exchange model respecting with the heavy quark spin symmetry where the parameter is obtained to reproduce the Tcc binding energy. We apply this model with superflavor symmetry to the D¯(*)Ξcc(*) molecule and predict a bound state with I(JP)=0(12−). If the pentaquark state corresponding to D¯(*)Ξcc(*) molecular state is observed in future experiments as predicted in this work, it is more likely that Tcc is a DD* molecular state. Published by the American Physical Society 2024

Ultrafast, Fano resonant colorimetric sensor with high chromaticity beyond standard RGB

Fast-responsive colorimetric sensors with a wide color gamut have garnered significant attention for real-time atmospheric monitoring observable to the naked eye. Although swelling medium-based Fabry–Perot cavities, which enable linear resonance shifts with high Q-factors, have been widely suggested, they face limitations such as a restricted color gamut within standard RGB due to subtractive colors and slow response times caused by the top layer blocking, delaying the swelling medium’s equilibrium time. Here, we present two-dimensionally nanostructured Fano resonant colorimetric sensors using a swelling medium with significantly improved responsiveness and color representation beyond standard RGB. The nanostructured Fano resonator is elaborately designed to transform the spectral line shape into a Lorentz state in terms of reflectance, resulting in additive color through controlled coupling parameters of the resonator systems. In addition, the nanostructuring of the surface provides direct channels to water vapors, ensuring fast and strong interaction with the swelling medium. Consequently, the fabricated sensor exhibits a wide color gamut, covering 141% of standard RGB and 105% of Adobe RGB, and demonstrates rapid responsiveness with response and recovery times of 287 ms and 87 ms, respectively.

Conjunctive representation of colors and shapes in human occipitotemporal and posterior parietal cortices

Conjunctive representation of colors and shapes in human occipitotemporal and posterior parietal cortices

Parallel gain modulation mechanisms set the resolution of color selectivity in human visual cortex.

Color discrimination is fundamental to human behavior. We find bananas by coarsely searching for yellow but then differentiate nuances of yellow to pick the best exemplars. How does the brain adjust the resolution of color selectivity to our changing needs? Here, we analyze the brain magnetic response in the human visual cortex to show that color selectivity is adaptively set by coarse- and fine-resolving processes running in parallel at different hierarchical levels. Those include a gain enhancement in the higher-lever cortex of color units tuned away from the target to resolve very similar colors and a coarsely resolving gain enhancement in the mid-level cortex of units tuned to the target. Our findings suggest that attention operates on a form of multiresolution representation of color at different levels in the visual hierarchy, which keeps selectivity adaptive to a changing resolution context.

Asymmetric distribution of color-opponent response types across mouse visual cortex supports superior color vision in the sky.

Color is an important visual feature that informs behavior, and the retinal basis for color vision has been studied across various vertebrate species. While many studies have investigated how color information is processed in visual brain areas of primate species, we have limited understanding of how it is organized beyond the retina in other species, including most dichromatic mammals. In this study, we systematically characterized how color is represented in the primary visual cortex (V1) of mice. Using large-scale neuronal recordings and a luminance and color noise stimulus, we found that more than a third of neurons in mouse V1 are color-opponent in their receptive field center, while the receptive field surround predominantly captures luminance contrast. Furthermore, we found that color-opponency is especially pronounced in posterior V1 that encodes the sky, matching the statistics of natural scenes experienced by mice. Using unsupervised clustering, we demonstrate that the asymmetry in color representations across cortex can be explained by an uneven distribution of green-On/UV-Off color-opponent response types that are represented in the upper visual field. Finally, a simple model with natural scene-inspired parametric stimuli shows that green-On/UV-Off color-opponent response types may enhance the detection of 'predatory'-like dark UV-objects in noisy daylight scenes. The results from this study highlight the relevance of color processing in the mouse visual system and contribute to our understanding of how color information is organized in the visual hierarchy across species.

ETBHD‐HMF: A Hierarchical Multimodal Fusion Architecture for Enhanced Text‐Based Hair Design

AbstractText‐based hair design (TBHD) represents an innovative approach that utilizes text instructions for crafting hairstyle and colour, renowned for its flexibility and scalability. However, enhancing TBHD algorithms to improve generation quality and editing accuracy remains a current research difficulty. One important reason is that existing models fall short in alignment and fusion designs. Therefore, we propose a new layered multimodal fusion network called ETBHD‐HMF, which decouples the input image and hair text information into layered hair colour and hairstyle representations. Within this network, the channel enhancement separation (CES) module is proposed to enhance important signals and suppress noise for text representation obtained from CLIP, thus improving generation quality. Based on this, we develop the weighted mapping fusion (WMF) sub‐networks for hair colour and hairstyle. This sub‐network applies the mapper operations to input image and text representations, acquiring joint information. The WMF then selectively merges image representation and joint information from various style layers using weighted operations, ultimately achieving fine‐grained hairstyle designs. Additionally, to enhance editing accuracy and quality, we design a modality alignment loss to refine and optimize the information transmission and integration of the network. The experimental results of applying the network to the CelebA‐HQ dataset demonstrate that our proposed model exhibits superior overall performance in terms of generation quality, visual realism, and editing accuracy. ETBHD‐HMF (27.8 PSNR, 0.864 IDS) outperformed HairCLIP (26.9 PSNR, 0.828 IDS), with a 3% higher PSNR and a 4% higher IDS.

Color Analysis of Brocade from the 4th to 8th Centuries Driven by Image-Based Matching Network Modeling

To achieve the color matching rules for the textiles discovered during Silk Road excavations between the 4th and 8th centuries, this research proposed an image-based matching network modeling method. The Silk Road facilitated trade and cultural exchange between the East and West, and the textiles found along the way depict the development of fabrics in a color scheme with great cultural significance. A total of 165 images with brocade patterns were collected from a book with a detailed description of the Western influences on textiles along the Silk Road. Two different clustering methods, including the K-means clustering method and octree quantization approach, were used to extract the primary and secondary colors. By combining the HSV color space with the PCCS color system, the color distribution was analyzed to discover the features of representative color patterns. The co-occurrence relationship of the auxiliary colors was explored using the Apriori algorithm, and a total of eight association rules were established. The results showed that the K-means clustering algorithm can show a better effect of color classification to obtain three primary colors and nine secondary colors. The matching mechanism with a visualized network model was also proposed, which showed that reddish-yellow tones are the main colors in the brocade patterns, and the light and soft tones separately account for 27% and 20%. Beige and brown are the most common colorways, with a confidence level of 47%. One style of brocade pattern was used to demonstrate different appearances within various color networks, which could be applied to 3D virtual fitting. This image-based matching network modeling approach makes the color matching schemes visible, and can assist fashion design with fabric features influenced by historical and cultural development.

53381 Skin Color Representation in Public Images used to Depict Conditions Commonly Affecting Skin of Color

53381 Skin Color Representation in Public Images used to Depict Conditions Commonly Affecting Skin of Color

53826 Skin of Color Representation in Drug Advertisements for Biologic Treatments FDA-approved for Psoriasis

53826 Skin of Color Representation in Drug Advertisements for Biologic Treatments FDA-approved for Psoriasis