Computer Color Research Articles

Illuminant segmentation for multi-illuminant scenes using latent illumination encoding

Color constancy is an important part of the Human Visual System that allows us to recognize colors of object invariant to the light that is illuminating them. Computational color constancy is the process of estimating the illumination of a scene using some computational method. However, this problem is inherently ill-posed. The RGB value of each pixel is dependent on spectral reflectance of the object and the spectral power distribution of the illumination. Hence, methods that try to solve the computational color constancy problem have to introduce assumptions about the illumination. One common assumption is that there is only one global source of illumination, i.e., that the illumination is constant across the whole scene. Under this assumption, modern color constancy methods achieve excellent results, usually predicting the illumination color with accuracy better than the human eye. This assumption is broken in many real-world multi-illuminant scenes, e.g., outdoor images where parts of the scene are illuminated by either sunlight or skylight. This leads to significant drop in accuracy of single-illuminant estimation methods. Therefore, in this work, we propose a novel method for segmenting images based on illumination in multi-illuminant scenes. This method detects regions where there is only one illuminant, thus detecting areas where the single-illuminant assumption holds. We show that our method produces excellent results and outperforms all baseline models by a large margin.

Research on Art Design System of Industrial Dyestuff Colour Matching with Computer Aid

The success rate of computer colour matching not only depends on the colour matching theory and algorithm, but also limited by the basic data of colour matching. In addition to the traditional accuracy analysis method, colour gamut analysis is added. We take the establishment of basic data for dyeing polyester fabrics with disperse dyes as an example to discuss the rationality of basic data for computer colour matching. The model can obtain colour matching results with a colour difference of less than 1 after 1-2 colour corrections, and can meet the basic requirements of multi-component, fixed-proportion colour matching and colour correction in the production of colour spinning.

Prediction of the reactive dye recipe in cotton fabric dyeing using the pad‐dry‐pad‐steam process by particle swarm optimisation‐least squares support vector machine with matching database

AbstractCompared with manual matching, computer colour matching is faster and more reliable. Computer colour‐matching results are determined by the method used; however, in practical application, the accuracy of colour matching still needs to be improved. In this paper, we first designed a series of colour‐matching schemes with two reactive dyes, Levafix Blue and Levafix Amber, then dyed cotton fabric through the pad‐dry‐pad‐steam process, and finally established the matching database. Furthermore, two colour‐matching sub‐models were developed by least squares support vector machine (LSSVM) to predict the dye recipe in cotton fabric dyeing, while particle swarm optimisation (PSO) was applied to optimise and tune the parameters of the LSSVM models. Herein, the model inputs are the colour parameters L*, a* and b* of the dyed samples and the model output is dye concentration. It is confirmed that the colour‐matching models have excellent evaluation indexes and colour‐matching effects. The coefficient of determination is more than 0.99, and the colour‐matching precision is more than 98%. All the results prove that the colour‐matching models by PSO‐LSSVM can be accurately applied in practical dyeing to predict the reactive dye recipe, which is suitable for use on an industrial scale.

Ramifications of incorrect image segmentations; emphasizing on the potential effects on deep learning methods failure

IntroductionDetecting failure cases is critical to ensure a secure self-driving system. Any flaw in the system directly results in an accident. In genuine class, the model’s probability reflects better-reflected model confidence. As a result, the confidence distributions of failed predictions were changed to lower values. In contrast, accurate predictions were remained associated with high values, allowing for considerably more excellent separability between such prediction types. The study investigates the association of ramifications with computational color constancy that can negatively influence CNN’s image classification and semantic segmentation.MethodologyImage datasets were used to conduct different scales and complexity experiments. For instance, minimal and straightforward images of digits were comparatively provided through MNIST and SVHN datasets. The dataset’s standard validation set was employed to test and compute additional metrics because ground truth that is not publicly available for some test sets.ResultsThe results depicted that baseline methods were outperformed through the proposed approach with a considerable variant on minimal datasets or models in every context. Therefore, Transmission Control Protocol (TCP) is appropriate in failure prediction, and ConfidNet is competent to be fulfilled as confidence criterion. Further, one of the solutions would be to elevate the validation set size, but this would influence the prediction performance of a failure model. On the contrary, the confidence estimation was based on models with test predictive performance levels, similar to baselines.ConclusionsThe gap between validation accuracy and training accuracy was significant on CIFAR-100, which indicates the modest enhancement for failure detection via the validation set.

A review on reverse micellar approach for natural fiber dyeing

AbstractThis study reviews the need and importance of non‐aqueous dyeing systems in the chemical processing industry, especially using the reverse micellar system in the dyeing of natural fibres. When it comes to conventional dyeing systems, which use large quantities of water, chemical, and energy, cause a lot of effluent load to the environment. Again the effluent water has to be treated well to eliminate all the harmful substances in it. To overcome the issues, a lot of research has been carried out in this area to minimise the use of water and chemicals in the dyeing process. Like use of low material to liquor ratio (M:L) in dyeing, dyes with high fixation at lower temperatures, low/salt less dyeing in case of reactive dyeing with cotton, cottonseed oil dyeing, microwave‐assisted dyeing, and use of supercritical carbon dioxide in case of polyester dyeing and many more. All the above said methods are aqueous‐based and, after the completion of dyeing, results in wastewater generation, which requires further treatment to reduce the harmful chemicals. Therefore, to further minimise the use of water and chemicals in the dyeing of natural fibres reverse micellar system has been introduced with the help of surfactants of both Ionic and nonionic in nature. As of today, a lot of work has been carried out in the dyeing of natural fibres with this system by employing ionic, nonionic, and mixed surfactants. Fascinating results were obtained in the dyeing with good levelness, high exhaustion, and fixation values, and results were compared with conventional dyeing. Computer colour matching studies were also done better to understand the applicability of these systems in the industry and found nearer results.

The 60th Anniversary of TFTs and the Evolution of High‐Resolution Displays

The 60th Anniversary of TFTs and the Evolution of High‐Resolution Displays

Which Features Are More Correlated to Illuminant Estimation: A Composite Substitute

Computational color constancy (CCC) is to endow computers or cameras with the capability to remove the color bias effect caused by different scene illuminations. The first procedure of CCC is illuminant estimation, i.e., to calculate the illuminant color for a given image scene. Recently, some methods directly mapping image features to illuminant estimation provide an effective and robust solution for this issue. Nevertheless, due to diverse image features, it is uncertain to select which features to model illuminant color. In this research, a series of artificial features weaved into a mapping-based illuminant estimation framework is extensively investigated. This framework employs a multi-model structure and integrates the functions of kernel-based fuzzy c-means (KFCM) clustering, non-negative least square regression (NLSR), and fuzzy weighting. By comparing the resulting performance of different features, the features more correlated to illuminant estimation are found in the candidate feature set. Furthermore, the composite features are designed to achieve the outstanding performances of illuminant estimation. Extensive experiments are performed on typical benchmark datasets and the effectiveness of the proposed method has been validated. The proposed method makes illuminant estimation an explicit transformation of suitable image features with regressed and fuzzy weights, which has significant potential for both competing performances and fast implementation against state-of-the-art methods.

Comparison of computer vision system and colour spectrophotometer for colour measurement of mozzarella cheese

The colour of food is one of the most important selection criteria affecting purchasing decision of a consumer. Conventionally, colour is measured using colorimeters and colour spectrophotometer. In recent times, computer vision technology has emerged as a robust technique and can be implemented as an alternative to conventional colour measurement methods. A cost-effective computer vision system was developed for measuring the colour of mozzarella cheese with provision to mount different cameras. Various CCD and CMOS based digital cameras were calibrated and evaluated for image acquisition of cheese samples. An additional validation step was introduced to characterize different image acquisition devices on the basis of Root Mean Square Deviation (RMSD) and Mean Normalized Error (MNE). The colour measurement algorithm extracts RGB values from the sample image and coverts into CIE L*a*b* units. The conversion between two colour spaces was done employing standard equations. The colour of mozzarella cheese samples was measured in the CIE L*a*b* color space with the computer vision system (CVS) and commercial colour spectrophotometer. Both the systems yielded equivalent results and developed CVS can be used to study colorimetric quality parameters of mozzarella cheese.

The color matching design based on polynomial regression

Camouflage spots are mostly color patches of different colors. Compared with the traditional empirical color matching method, computer color matching can quickly and effectively calculate the proportion of red, yellow, and blue primary colors. In general, camouflage computer color matching in the practical application of the initial formula is not accurate and multiple color modification still cannot match. First, we measured the chroma coordinates of 98 standard swatches made of red, green, and blue coatings in different proportions by the OHSP-660 spectral reflectance tester. Then, we adopted polynomial regression algorithm by Statistical Product and Service Solutions (SPSS) to obtain the regression equation between [Formula: see text], [Formula: see text], [Formula: see text], and [Formula: see text]. Finally, the fitting equation is used to obtain the paint color scheme in the coordinates of [Formula: see text], [Formula: see text], and [Formula: see text] in arbitrary uniform color space, so as to realize the purpose of computer color matching. The experimental results show that the minimum color difference is 1.691 ( L* a* b* unit) the maximum color difference is 2.497, which meets the design requirements of camouflage and target color difference being no more than 3. Our work provides a theoretical basis for the further development of computer color matching system in camouflage.

Illuminant Estimation Using Adaptive Neuro-Fuzzy Inference System

Computational color constancy (CCC) is a fundamental prerequisite for many computer vision tasks. The key of CCC is to estimate illuminant color so that the image of a scene under varying illumination can be normalized to an image under the canonical illumination. As a type of solution, combination algorithms generally try to reach better illuminant estimation by weighting other unitary algorithms for a given image. However, due to the diversity of image features, applying the same weighting combination strategy to different images might result in unsound illuminant estimation. To address this problem, this study provides an effective option. A two-step strategy is first employed to cluster the training images, then for each cluster, ANFIS (adaptive neuro-network fuzzy inference system) models are effectively trained to map image features to illuminant color. While giving a test image, the fuzzy weights measuring what degrees the image belonging to each cluster are calculated, thus a reliable illuminant estimation will be obtained by weighting all ANFIS predictions. The proposed method allows illuminant estimation to be dynamic combinations of initial illumination estimates from some unitary algorithms, relying on the powerful learning and reasoning capabilities of ANFIS. Extensive experiments on typical benchmark datasets demonstrate the effectiveness of the proposed approach. In addition, although there is an initial observation that some learning-based methods outperform even the most carefully designed and tested combinations of statistical and fuzzy inference systems, the proposed method is good practice for illuminant estimation considering fuzzy inference eases to implement in imaging signal processors with if-then rules and low computation efforts.

On the evaluation of temporal and spatial stability of color constancy algorithms.

Computational color constancy algorithms are commonly evaluated only through angular error analysis on annotated datasets of static images. The widespread use of videos in consumer devices motivated us to define a richer methodology for color constancy evaluation. To this extent, temporal and spatial stability are defined here to determine the degree of sensitivity of color constancy algorithms to variations in the scene that do not depend on the illuminant source, such as moving subjects or a moving camera. Our evaluation methodology is applied to compare several color constancy algorithms on stable sequences belonging to the Gray Ball and Burst Color Constancy video datasets. The stable sequences, identified using a general-purpose procedure, are made available for public download to encourage future research. Our investigation proves the importance of evaluating color constancy algorithms according to multiple metrics, instead of angular error alone. For example, the popular fully convolutional color constancy with confidence-weighted pooling algorithm is consistently the best performing solution for error evaluation, but it is often surpassed in terms of stability by the traditional gray edge algorithm, and by the more recent sensor-independent illumination estimation algorithm.

The effect of colour matching on perceptual integration of pictures and frames

ABSTRACT A black frame presented around one of 12 pictures during rapid serial visual presentation (RSVP) was frequently misperceived as surrounding the preceding or following picture in the sequence (temporal migration; Intraub, 1985, 1989). Would temporal migration of the frame be reduced if the frame and picture shared a common property (colour)? Two different colour frames were created for each picture: one matching and one mismatching the picture’s most prevalent colour (Experiment 1) as determined by a computational image colour summarizer (Krzywinski, 2018). Presentation rate was 100 ms/picture. Colour-matching significantly reduced temporal migration (4.3%), but the effect was small. To determine if greater colour-overlap would yield a larger effect, frames composed of colour “tiles” that matched or mismatched each picture’s four most prevalent colours were presented at a rate of 111 ms/picture (Experiment 2). Again, a small, significant reduction in temporal migration was observed (3.2%). Although pictures and frames are distinct forms, the probability of correct integration of the two appears to be affected by the presence of a common property (colour). Results are discussed in terms of an integrative short-term buffer. Possible application of the temporal migration paradigm to studying visual integration of components of real-world scenes is addressed.

Advancing Equity in Graduate Pathways: Examining the Factors that Sustain and Develop Computing Graduate Aspirations

ABSTRACT In light of the need to both grow and diversify graduate enrollments in computing fields, this quantitative study examined variables associated with students’ initial aspirations for computing graduate school as well as how aspirations were sustained or developed over time. Using a longitudinal sample of undergraduate students who enrolled in introductory computing courses at 15 research universities, findings suggest that environments and interactions that sustain undergraduate students’ initial computing graduate aspirations are quite different than those that contribute to other students’ development of graduate aspirations. Specifically, this study highlights the key role of domain-specific psychosocial beliefs (i.e., computing identity, computing self-efficacy) in shaping students’ aspirations for graduate computing. Results also highlight important equity issues, particularly related to supporting graduate aspirations among underrepresented Students of Color in computing. We offer implications for cultivating gender and racial/ethnic diversity in computing graduate school pathways, which is imperative to generating more inclusive computing environments in the tech industry and academia.

Investigation for the color samples with equi‐depth in Natural Color System

AbstractThe color depth, an important attribute of color, can reflect the amount of dye partly, which has important functions on the evaluation of color‐fastness and strength of dyes, dyeing effect of fabric, computer color matching, and so on. Natural Color System, an internationally accepted color system, orders colors by three parameters (blackness, chromaticness, and hue). The color depth has not been specified within the Natural Color System. This article tries to find the regularity of the sample with equal‐depth in Natural Color System. Firstly, 1950 color samples in Natural Color System were measured using an X‐Rite Color i7 Spectrophotometer, and their color depths were calculated by five color depth formulas. Then, trend analysis and mathematical modeling methods were used to achieve the connection between the color depth and the notations of Natural Color System basing on color depth theories. Results show that, in Natural Color System, the color samples with the same distance to pure white do not have equal depth; but the color samples with the same nuance (equal blackness, whiteness and chromaticness) have broadly equal color depth, and their average coefficient values are lower than that of Society of Dyers and Colourists Standard Depths. Besides, regressive formulas were built, with which the color depths of any chips in Natural Color System can be calculated broadly by their notation.

Newly Identified Hemodynamic Parameter to Predict Thin-Walled Regions of Unruptured Cerebral Aneurysms Using Computational Fluid Dynamics Analysis

The thin-walled regions (TIWRs) of intracranial aneurysms have a high risk of rupture during surgical manipulation. They have been reported to be predicted by wall shear stress and pressure (PS) based on computational fluid dynamics analysis, although this remains controversial. In this study, we investigated whether the oscillatory shear index (OSI) can predict TIWRs. Twenty-five unruptured aneurysms were retrospectively analyzed; the position and orientation of the computational fluid dynamics color maps were adjusted to match the intraoperative micrographs. The red area on the aneurysm wall was defined as TIWR, and if most of the regions on the color map corresponding to TIWR were OSI low (lower quartile range), time-averaged wall shear stress (TAWSS) high, or PS high (upper quartile range), each region was defined as a matched region and divided by the total number of TIWRs to calculate the match rate. In addition, the mean values of OSI, TAWSS, and PS corresponding to TIWRs were quantitatively compared with those in adjacent thick-walled regions. Among 27 TIWRs of 25 aneurysms, 23, 10, and 14 regions had low OSI, high TAWSS, and high PS regions (match rate: 85.2%, 37.0%, and 51.9%), respectively. Receiver operating characteristic curve analysis demonstrated that OSI was the most effective hemodynamic parameter (area under the curve, 0.881), followed by TAWSS (0.798). Multivariate analysis showed that OSI was a significant independent predictor of TIWRs (odds ratio, 18.30 [95% CI, 3.2800-102.00], P < 0.001). OSI may be a unique predictor for TIWRs. Low OSI strongly corresponds with TIWRs of intracranial aneurysms.

Evaluating the effect of light exposure and seed coat on lentil cotyledon color by computer vision

AbstractA computer vision system and color analysis algorithms were employed to study the influence of UVA (ultraviolet‐A) and visible light on the color of lentils with red, green, and yellow cotyledons. Twenty samples of cotyledons from each of the three‐color classes were subjected to six light treatments (ultraviolet, full‐spectrum visible, red, green, blue, and dark control) for 7 days at room temperature. The International Commission on Illumination L*a*b* (CIE L*a*b*) color values of the individual seeds were obtained before and after each treatment using the computer vision and image analysis system. Results of the analysis showed that light exposure had a statistically significant effect on all three cotyledon color classes. The effect size was largest for green lentils, smaller in yellow, and least in red lentils. By having established that light exposure affects the color of lentil cotyledons, it was hypothesized that seed coats may protect cotyledons against the effects of light exposure and that the degree of protection would vary with seed coat color classification. This hypothesis was tested using green‐cotyledon lentil varieties with different seed coat classes. Results confirmed that light‐induced color loss in the cotyledons was significantly influenced by seed‐coat color class. The order of protective effect of lentil seed coat from least to highest was found to be as follows: gray‐zero tannin, green, normal gray, and black. Thus, breeding for seed coat protection may improve the overall quality of green lentils. The results from this study will be informative to breeding programs that focus on enhancing the cotyledon color of lentils, and in making decisions regarding the dehulling of lentils and the design of dehulled lentil materials handling.

Illumination estimation challenge: The experience of the first 2 years

AbstractIllumination estimation is the essential step of computational color constancy, one of the core parts of various image processing pipelines of modern digital cameras. Having an accurate and reliable illumination estimation is important for reducing the illumination influence on the image colors. To motivate the generation of new ideas and the development of new algorithms in this field, two challenges on illumination estimation were conducted. The main advantage of testing a method on a challenge over testing it on some of the known datasets is the fact that the ground‐truth illuminations for the challenge test images are unknown up until the results have been submitted, which prevents any potential hyperparameter tuning that may be biased. The First illumination estimation challenge (IEC#1) had only a single task, global illumination estimation. The second illumination estimation challenge (IEC#2) was enriched with two additional tracks that encompassed indoor and two‐illuminant illumination estimation. Other main features of it are a new large dataset of images (about 5000) taken with the same camera sensor model, a manual markup accompanying each image, diverse content with scenes taken in numerous countries under a huge variety of illuminations extracted by using the SpyderCube calibration object, and a contest‐like markup for the images from the Cube++ dataset. This article focuses on the description of the past two challenges, algorithms which won in each track, and the conclusions that were drawn based on the results obtained during the first and second challenge that can be useful for similar future developments.

Similarity Evaluation on the Compound TCM Formulation "Huoling Shengji Granule" and Its Placebo by Intelligent Sensory Evaluation Technologies and the Human Sensory Evaluation Method Based on Critical Quality Attributes.

To evaluate the similarity of Huoling Shengji granule (HLG) and its placebo at both granules and solution status, the innovative methods that consist of intelligent sensory evaluation technologies and human sensory evaluation methods were developed based on critical quality attributes (CQAs) of granule. The CQAs for traditional Chinese medicine (TCM) placebo granule were mainly divided into three categories: formulation attributes, visual attributes, and attributes of taste and smell. In this investigation, the novel intelligent sensory evaluation technologies including the physical property testing apparatus, computer vision system, color card, and electronic tongue (E-tongue) were employed for characterization of CQAs of HLG and its placebo. Meanwhile, human sensory evaluation by test panels was used to description the HLG and its placebo in terms of appearance, color, taste, and smell. On that basis, the similarity of placebo to CQAs of HLG was assessed by calculating the angle cosine values. The intelligent and human sensory evaluation results showed that the similarity values of HLG and its placebo about the CQAs at granule and solution status were all close to 1, which means that the two preparations have high similarities. In this study, the established similarity evaluation methods based on the CQAs were convenient and reliable, which can be utilized to evaluate the similarity of TCM granule and their placebo at granule and solution status, and demonstrated to be well applied in placebo-controlled trials.

Dyeing Wool Knitted Fabric in Nano-scale Reverse Micelle with Reactive Dyes — A Computer Colour Matching Study

The colour quality assessment of reverse micelle-based (in octane or nonane medium) and conventional water-based dyed wool knitted fabric has been investigated using computer colour matching (CCM) approach. Both water-based and non-aqueous based dyeing methods exhibited no chromatic shift from the reflectance curves and established good linearity in the construction of calibration curves. The linearity of both calibration curves in terms of R2 value played a critical role in CCM. The colour yield obtained from reverse micellar dyed fabric samples was higher than the colour yield values from wool knitted fabrics dyed in water bath under different dye percentages. The CCM predicted concentrations were in good agreement to theoretical concentrations for both dyeing systems. To conclude, the results showed that reverse micellar dyeing in octane and nonane is capable of generating reliable colour recipes. The relative unlevelness indices (RUI) obtained from dyed wool knitted fabrics using reverse micellar approach reveal good to excellent levelness, which fulfills the requirements of the textile processing industry. TEM picture illustrated that reverse micelles are in nano-scale and this could enhance the dyeing effect of wool with reactive dyes.

In color constancy: data mattered more than network

The objective of this paper is to argue that data mattered more than network in terms of color constancy. Computational color constancy is a linear operation device-dependent, which is part of the camera imaging pipeline. We extend the dataset based on this pipeline and prove that the scene illumination can be predicted using a very simple network as long as the dataset is large enough and evenly distributed. In the process of expanding the dataset, firstly, we remove illumination color casts in images which is ground-truth illumination color and then casts randomly generated evenly distributed illumination colors in images. We randomly generate five labels for each image and then work on the image to obtain this dataset. Using this dataset, we introduce a very simple network that is able to compute the color mapping function to correct the image’s colors. Experiments on our new datasets demonstrate that the method of this paper significantly outperforms the state-of-the-art color constancy methods.