Search

Many fruit recognition works have applied statistical approaches to make an exact correlation between low-level visual feature information and high-level semantic concepts given by predefined text caption or keywords. Two common fruit recognition models include bag-of-features (BoF) and convolutional neural network (ConvNet), which achieve high-performance results. In most cases, the overfitting problem is unavoidable. This problem makes it difficult to generalize new instances with only a slightly different appearance, although belonging to the same category. This article proposes a new fruit recognition model by associating an object's low-level features in an image with a high-level concept. We define a perceptual color for each fruit species to construct a relationship between fruit color and semantic color name. Furthermore, we develop our model by integrating the perceptual color and semantic template concept to solve the overfitting problem. The semantic template concept as a mapping between the high-level concept and the low-level visual feature is adopted in this model. The experiment was conducted on three different fruit image datasets, with one dataset as train data and the two others as test data. The experimental results demonstrate that the proposed model, called perceptual color on semantic template (PCoST), is significantly better than the BoF and ConvNet models in reducing the overfitting problem.

It is well known that different names of color can lead to distinct attractions to people. To study the neural mechanism underlying this phenomenon, an implicit association test task was designed for color names, in which participants were required to select the possible meanings of a Greek phrase from two color names (in Chinese). The behavioral results showed that the participants were more likely to select novel names for long Greek phrases and dates names for short Greek phrases. The EEG results showed that the mean amplitude of N1 was greater for selections of novel color names than selections of dates names for Greek phrases. Meanwhile, the mean amplitude of N3 for novel color names was more negative than that of dates color names. Significant interaction effect of N3 was also found for the four kinds of selections between Greek phrases and Chinese color names. Moreover, a frontal-positive and occipital-negative distribution for scalp topography of N1 was found, while the scalp topography of N3 was opposite as frontal-negative and occipital-positive distribution, suggesting the importance of visual cortex for perception of the color names and prefrontal cortex for integration and decision of selection. In summary, the results here indicated that colors with novel names could easily attract people’s attention than colors with dates names, which might shed light on the usage of color names in real life. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1007/s11571-021-09685-y) contains supplementary material, which is available to authorized users.

The potential interaction between color naming and psychophysical color recognition has been historically debated. To study this interaction, here we utilized two approaches based on individual differences in color naming and variation of color name density along the color wheel. We tested a pool of Persian speaking subjects with a simple color matching task under two conditions: perceptual and memory-based matching. We also asked subjects to freely name 100 evenly sampled hues along the color wheel. We found that, individuals who possess more names to describe the color wheel have a strong edge in color memorization over those with fewer names. Nevertheless, having more or fewer color names was not related to the subjects’ performance in perceptual color matching. We also calculated the density of color names along the color wheel and observed that parts of the color wheel with higher density of color names are held in memory more accurately. However, similar to the case of individual differences, the density of color names along the wheel did not show any correlation with perceptual color matching performance. Our results demonstrate a strong link between color naming and color memorization both across different individuals and different parts of the color wheel. These results also show that low-level perceptual color matching is not related to color naming, suggesting that the variation in color naming—among the individuals and across the color wheel—is neither the cause nor the effect of variation in low-level color perception.

In this paper, we will investigate the contribution of color names for the task of salient object detection. An input image is first converted to color name space, which is consisted of 11 probabilistic channels. By exploiting a surroundedness cue, we obtain a saliency map through a linear combination of a set of sequential attention maps. To overcome the limitation of only using the surroundedness cue, two global cues with respect to color names are invoked to guide the computation of a weighted saliency map. Finally, we integrate the above two saliency maps into a unified framework to generate the final result. In addition, an improved post-processing procedure is introduced to effectively suppress image backgrounds while uniformly highlight salient objects. Experimental results show that the proposed model produces more accurate saliency maps and performs well against twenty-one saliency models in terms of three evaluation metrics on three public data sets., http://www.loujing.com/cns-sod/

The technology of visual object tracking based on correlation filter has good accuracy and efficiency. However, it is still necessary to be study further on the appearance model of the target, the scale variation of the target and so on. This paper proposes a tracking algorithm based on discriminative correlation filtering and a hybrid color feature. The hybrid color feature is composed of two parts, which are compressed color name features and Histogram of Oriented Gradient features based on opponent color space. These two parts features above are extracted from the target patch, respectively. For the first part, color-name features are extracted from a target patch firstly, and then block-based compressed color-name features are extracted according to these color-name features. For the second part, opponent color features are extracted from the target patch firstly, and then HOG features are extracted from these opponent color features. At the basis of the hybrid color feature, two different discriminative correlation filters are used to estimate the translation and the scale of the target, respectively. Finally, extensive experiments show that the tracking algorithm with the hybrid color features of this paper outperforming several state-of-the-art tracking algorithms.

"Yuanlianglv(forgiveness-green)" is a neologism about color, its referential meaning refers to green, while its cultural meaning is a kind of joke about being cuckolded. "Yuanlianglv" is not only a compound word but also a newly-emerging color word and a buzzword. According to its triple attributes, this paper expounds on the meaning construction of "yuanlianglv" by conceptual blending theory and analyzes the differences between "yuanlianglv" and common compound color words. Besides, this paper concludes four characteristics of color buzz words: time-sensitivity of sticking to the network hotspots, the internal driving force of socio-cultural background, the social dialect showing group emotions, and harmonious rhythm of syllables.

There are a variety of factors on stage environment such as dramatically changing lights, mutation and similar appearance of performers that make computer vision based performer detection task challenging. Relighting technique is adopted to handle the changing lights problem and can greatly relieve the negative effects caused by changing lights on detection results. Color transfer is one of the commonly used relighting techniques in performer detection task. However, existing color transfer methods will bring artifacts like structural defects, distortions and chromatic noises to processed images that make detection task fail. In this paper, we propose a novel color transfer network to improve the detection results. The proposed network is designed to switch the color distributions between two images. It can either transfers the color style of stage performance to public detection datasets to greatly expand the training datasets, or normalizes the style of captured stage images to some standard template to stable the data distribution. The proposed color transfer network deals with color transfer task in a one-shot framework, i.e. consistently training and predicting, taking advantage of the color name regularization. The effectiveness of the proposed method is validated by experiments on real-captured stage datasets.

A study of over seventy African languages suggests that Africa displays a strongly areal distribution of different “basic color term” (BCT) systems, and there is obvious impact of salient cultural activities, language contact, urbanization, and education on the evolution of color term systems. The smallest systems, containing BCTs for ‘red’, ‘black’, and white’, are predominantly in western Africa, though some East African Nilo-Saharan languages also appear to have this system. Most West African languages fall into Greenberg’s Niger-Congo phylum, but there is strong debate about whether these are a single genetic group. Southern African “Khoisan” and southern Narrow Bantu often have moderately elaborated systems with around five probable BCTs. The greatest elaboration and variation occur in Afro-Asiatic and some Nilo-Saharan languages of north and eastern Africa, with up to seven or eight BCTs plus salient “color+pattern” vocabulary.

In this paper, we study to name a color of an object in the given image, where the number of color names is 11 English basic color names, including pink, purple, red, orange, yellow, green, blue, gray, brown, black, and white. Our goal is to apply machine learning algorithms for assigning a color name to a principal object in images. Here, we evaluate three basic classification algorithms, namely Support Vector Machine (SVM), naive Bayes, and k-Nearest Neighbor (k-NN). The feature vector of the object in an image is the pixel intensities represented in different color spaces including RGB, HSV, Lab and CYMK. We randomly collect the training pixel intensities with their corresponding color names from the images obtained from the Internet. We then evaluate three classification algorithms using the e-Bay dataset. The experimental results show that the linear SVM with Lab color space is the best choice for solving this task.

Human color vision is a kind of subjective perception. Individual observers may have different color visions. However, the widely used RGB values were highly correlated with the display devices and even the same RGB value represented different colors on various display devices. Furthermore, physically identical patches with identical chromaticity coordinate values had different color perceptions in different spatial contexts. At the same time, individual color vision did not prevent people from color-related communications through color vocabulary. Considering the above problems and the characteristics of color vocabulary used in daily life for communication, we built a color vocabulary labeled data set based on human subjective color visual perception through corresponding visual psychophysics experiments, and then used the data set to train deep neural networks. The network can automatically identify 11 Chinese color vocabularies corresponding to images. The indicator F1-measure of the model reached an accuracy rate of more than 75% in different test sets. Based on the subjective visual perception data set and the deep neural network, we tried to avoid the perceived difference of other color extraction methods.

We propose an explainable model for classifying the color of pixels in images. We propose a method based on binary search trees and a large peer-labeled color name dataset, allowing us to synthesize the average human perception of colors. We test our method on the application of Person Search. In this context, persons are described from their semantic parts, such as hat, shirt, ... and person search consists in looking for people based on these descriptions. We label segments of pedestrians with their associated colors and evaluate our solution on datasets such as PCN and Colorful-Fashion. We show a precision as high as 83% as well as the model ability to generalize to multiple domains with no retraining.

In this paper, we introduce formal definitions of the concepts of fuzzy color and fuzzy color space. First, we formalize the notion of fuzzy color for representing the correspondence between computational representation of colors and perceptual color categories identified by a color name. Second, we propose a methodology for learning fuzzy colors based on the paradigm of conceptual spaces, where prototypes are used for each category to be learnt. Since the conceptual space approach yields crisp categorizations, we introduce a novel methodology for defining fuzzy boundaries of color categories on the basis of a Voronoi tessellation of a color space. Finally, we also formalize the notion of fuzzy color space as the collection of fuzzy colors corresponding to the color categories employed in a certain context/application and/or for a specific user. Different typologies of fuzzy color spaces are proposed in order to be consistent with the nature of the categories we want to model. Our approach is illustrated by defining fuzzy color spaces using RGB with the Euclidean distance. Examples based on the well-known ISCC-NBS color naming system are presented, as well as others based on collections of color names and prototypes provided by users. The proposal is evaluated and compared with the most used approaches for color modeling. Additionally, a website located at http://www.jfcssoftware.com including all experimentation data, software implementing our models, and additional materials is available to researchers in color modeling.

In grapheme-color synesthesia, seeing a particular letter or number evokes the experience of a highly specific color. Here we investigate the brain’s real-time processing of words in this population, by recording event-related brain potentials (ERPs) from 15 grapheme-color synesthetes and 15 controls as they judged the validity of word pairs (“yellow banana” vs. “blue banana”) presented under high and low visual contrast. Relative to high contrast words, low contrast words elicited ~30ms delayed P1/N170 visual ERP components in both groups. When color concepts were conveyed to synesthetes by a string of achromatic graphemes (“55555 banana”) individually chosen for each synesthete, visual contrast effects were like those in color words: P1/N170 components were delayed 30ms, but unchanged in amplitude. When controls saw equivalent strings of colored graphemes, visual contrast modulated the amplitude of P1/N170, leaving their latency unchanged. Color induction in synesthetes thus differs from color perception in controls. Independent from the experimental effects, all orthographic stimuli elicited larger N170 and P2 in synesthetes than controls. While P2 (150–250ms) enhancement was similar in all synesthetes, N170 (130–210ms) amplitude varied with the subjective experience of synesthetes and their self-reported imagery ability. Results suggest the extent of immediate cross-activation in visual areas processing color and shape is most pronounced in so-called projector synesthetes whose concurrent colors are experienced as originating in external space.

The article examines the use of the color brown in Old Norse-Icelandic literature, which is encoded bybrúnnandjarpr. More specifically, it seeks to determine through linguistic categorization the objects about which brown is used and to determine on the basis of its frequency whether for Old Norse-Icelandic brown should be placed in the earlier stages of the evolution of color terms or if it should be assigned to the later stages. The data show thatbrúnnis the more frequently used term, though the earliest texts suggest that bothbrúnnandjarprwere contextually restricted. Gradually,brúnncame to be applied to a wider range of objects, whereasjarprremained a secondary color term. As a basic color term,brúnnshould be assigned a fairly late stage in the temporal-evolutionary order of basic color terms.

Color terms are used to express light spectrum characteristics captured by human vision, and color naming across languages partition color spaces differently. Such partition differences have been surveyed through several empirical experiments that employ Munsell color chips. We propose a novel visualization method for color terms based on thousands of images collected from query results provided by an image search engines such as Google. A series of experiments was conducted using eight basic color terms in seven languages. Pixel values in the images are counted to form color histograms according to the color pallet used in the world color survey. The visualization results can be summarized as follows: (1) Japanese and Korean color terms have wider distributions in the color space than terms in other languages do and (2) color visualizations for color terms pink and brown are affected by their links to proper nouns. Highlights Color terms are visualized with 3D histograms on WCS pallet. Color frequencies are assembled automatically from internet images. Red, orange, and yellow have sharp peaks. French pink has several peaks. Pink and brown are affected by their links to proper nouns. English and French colors have smaller distributions than those of Japanese and Korean.

Aiming at the problem of poor robustness and the low effectiveness of target tracking in complex scenes by using single color features, an object-tracking algorithm based on dual color feature fusion via dimension reduction is proposed, according to the Correlation Filter (CF)-based tracking framework. First, Color Name (CN) feature and Color Histogram (CH) feature extraction are respectively performed on the input image, and then the template and the candidate region are correlated by the CF-based methods, and the CH response and CN response of the target region are obtained, respectively. A self-adaptive feature fusion strategy is proposed to linearly fuse the CH response and the CN response to obtain a dual color feature response with global color distribution information and main color information. Finally, the position of the target is estimated, based on the fused response map, with the maximum of the fused response map corresponding to the estimated target position. The proposed method is based on fusion in the framework of the Staple algorithm, and dimension reduction by Principal Component Analysis (PCA) on the scale, the complexity of the algorithm is reduced, and the tracking performance is further improved. Experimental results on quantitative and qualitative evaluations on challenging benchmark sequences show that the proposed algorithm has better tracking accuracy and robustness than other state-of-the-art tracking algorithms in complex scenarios.

The naming of colors has long been a topic of interest in the study of human culture and cognition. Color term research has asked diverse questions about thought and communication, but no previous research has used an evolutionary framework. We show that there is broad support for the most influential theory of color term development (that most strongly represented by Berlin and Kay [Berlin B, Kay P (1969) (Univ of California Press, Berkeley, CA)]); however, we find extensive evidence for the loss (as well as gain) of color terms. We find alternative trajectories of color term evolution beyond those considered in the standard theories. These results not only refine our knowledge of how humans lexicalize the color space and how the systems change over time; they illustrate the promise of phylogenetic methods within the domain of cognitive science, and they show how language change interacts with human perception.

To investigate linguistic and perceptual boundaries within the “blue” region of the color gamut, we analyzed sorting data from speakers of six languages who sorted color stimuli by similarity. Two of these languages, Russian and Italian, are thought to have a separate basic color term and category for “light blue,” distinguishing it from “blue,” and the third was English, which lacks this distinction. There has been less study of the possible basicness of “light blue” terms in the other three languages: Lithuanian and Estonian (both spoken in Baltic states) and Udmurt (a linguistic enclave in North Russia). Sorting data from each sample of speakers were analyzed with principal component analysis and multidimensional scaling, reducing them to a pattern of interstimulus similarities. In addition, color-naming data were collected for five of the languages and confirmed that sorting responses were not simply a reflection of the words used by subjects to label the stimuli. A “clustering index” was created, quantifying the extent of light/dark blue separation and the strength of any category boundary between them; this was low for English-language participants but high for Russian and Italian. Udmurt and Lithuanian values were also high, whereas Estonian responses were closer to English. Thus, when clustering of blue stimuli is used as an additional indicator of basicness, the results are compatible with earlier evidence that “light blue” is a separate basic color category in Russian and Italian, and further indicate that light blue terms are basic in Udmurt and Lithuanian, but not Estonian. It may be that “blue” categories are especially susceptible to splitting into two under the influence of linguistic contact. © 2016 Wiley Periodicals, Inc. Col Res Appl, 2016

We investigated the nature of covertly processed visual elements in a patient with simultanagnosia, a disorder characterized by the inability to perceive multiple aspects of a visual scene all at once. Using the first letter of the color words red, green, or blue, we created a novel testing paradigm that combined Navon global-local stimuli with a single-letter Stroop task. The letters R, G, or B were arranged in the overall configuration of a large R, G, or B. The patient never could report seeing the larger letter, and always could name the smaller letters. But, when asked to name ink color only, and ignore letter identity, the large letter covertly affected responding. That is, when the large letter was the same as the first letter of the ink color, the patient named ink color more quickly and accurately than when the large letter was incongruent with the correct response. Moreover, when the covert global and overt local visual processing conflicted, the global letter always dominated over the local letters, despite the patient's inability to perceive it consciously. These data show that the covert processing of global visual information in simultanagnosia can dominate overt local information, even across different streams of information processing.

When data categories have strong color associations, it is useful to use these semantically meaningful concept-color associations in data visualizations. In this paper, we explore how linguistic information about the terms defining the data can be used to generate semantically meaningful colors. To do this effectively, we need first to establish that a term has a strong semantic color association, then discover which color or colors express it. Using co-occurrence measures of color name frequencies from Google n-grams, we define a measure for colorability that describes how strongly associated a given term is to any of a set of basic color terms. We then show how this colorability score can be used with additional semantic analysis to rank and retrieve a representative color from Google Images. Alternatively, we use symbolic relationships defined by WordNet to select identity colors for categories such as countries or brands. To create visually distinct color palettes, we use k-means clustering to create visually distinct sets, iteratively reassigning terms with multiple basic color associations as needed. This can be additionally constrained to use colors only in a predefined palette.

The aim of the research reported by this study was on the one hand to identify what colors were associated with particular words in relation to a specific language (Italian), by portraying them in color stimuli on the screen of a monitor; and on the other hand to verify whether some words of that language denoted colors that were either particularly well defined or confused with others. In an experiment using special software, the subjects were asked to produce colors directly, instead of choosing among a number of colors presented on the screen. The results showed that (i) it is possible to identify the color-stimuli to which the terms of a language refer; that (ii) the “best” colors Giallo (Yellow), Rosso (Red), Blu (Blue), and Verde (Green) which the subjects were requested to produce were very similar to the corresponding unique hues; that (iii) among the mixed hues there were perceptually intermediate colors, that is, ones exactly midway between two consecutive unique colors: Arancione (Orange) and Viola (bluish Purple); that (iv) Turquoise and Lime were clearly positioned in the mental space of color of the participants; and that (v) for Italian speakers some hues coincide: Azzurro (Azure) and Celeste (Cerulean); Arancione (Orange), RossoGiallo (RedYellow) and Carota (Carrot); Lime and GialloVerde (YellowGreen), so that their color terms can be considered synonyms. Our most interesting finding, however, is that for Italian speakers these four mixed colors with their specific names (Lime, Turchese (Turquoise), Viola (bluish Purple) and Arancione (Orange) fall perceptually in the middle of each of the four quadrants formed in the hue circle by the four unique hues. The resulting circle is therefore characterized by eight colors of which four are unique and four are intermediate mixed. It would be advisable to repeat the study cross-culturally to test for possible similarities and differences in color meanings with speakers of different languages. © 2016 Wiley Periodicals, Inc. Col Res Appl, 42, 89–101, 2017

Competing theories attempt to explain the function of Broca's area in single word processing. Studies have found the region to be more active during processing of pseudo words than real words and during infrequent words relative to frequent words and during Stroop (incongruent) color words compared to Non-Stroop (congruent) words. Two related theories explain these findings as reflecting either “cognitive control” processing in the face of conflicting input or a linguistic prediction error signal, based on a predictive coding approach. The latter implies that processing cost refers to violations of expectations based on the statistical distributions of input. In this fMRI experiment we attempted to disentangle single word processing cost originating from cognitive conflict and that stemming from predictive expectation violation. Participants (N = 49) responded to whether the words “GREEN” or “RED” were displayed in green or red (incongruent vs congruent colors). One of the colors, however, was presented three times as often as the other, making it possible to study both congruency and frequency effects independently. Auditory stimuli saying “GREEN” or “RED” had the same distribution, making it possible to study frequency effects across modalities. We found significant behavioral effects of both incongruency and frequency. A significant effect (p < .05 FWE) of incongruency was found in Broca's region, but no effect of frequency was observed and no interaction. Conjoined effects of incongruency and frequency were found in parietal regions as well as in the Visual Word Form Area (VWFA). No interaction between perceptual modality and frequency was found in VWFA suggesting that the region is not strictly visual. These findings speak against a strong version of the prediction error processing hypothesis in Broca's region. They support the idea that prediction error processes in the intermediate timeframe are allocated to more posterior parts of the brain.

Previous claims of the number of color categories and corresponding basic color terms in modern Mandarin Chinese remain irreconcilable, mainly due to the shortage in objectively evaluating the basicness of color terms with statistical significance. Therefore the present study applied k-means cluster analysis to investigate native Mandarin Chinese speakers' color naming data of 330 color chips similar to those used in World Color Survey. Results confirmed that there are 11 basic color categories among modern Mandarin speakers in Taiwan, one corresponding to each basic color term. Results also showed that observers overwhelmingly agreed in their use of Mandarin color terms, including those that had yielded ambiguous results in previous studies (gray, brown, pink, and orange). There is significant cross-language similarity when comparing the distribution of color categories in the World Color Survey chart with American English and Japanese data. The motif analysis and group mutual information analysis suggest that Mandarin color terms used in Taiwan describe very similar categories and are, hence, similarly precise in communicating color information as those in Japanese and American English. These results show that three languages of fundamentally different cultures and histories have very similar basic color terms.

Certain modified color terms encountered in a multi-language corpus of unconstrained color-naming data, elicited with 65 Color-aid Corporation tiles, can be glossed into English as “bright (or vivid) X” (e.g., Estonian “ere-X”), while other modifiers are glossed “light” or “dark.” However, translation between languages or into the terms of colorimetry is never assured. We address the problem empirically by examining the denotata of each modified term and treating its uses as a distribution across a metric color space in which tiles are located as points. We compared each distribution with that of the unmodified term X, identifying the latter with the focus of X (the within-language consensus about the most prototypical exemplar of X). In some cases the modifiers operate as “bright” in the sense of “intense” or “saturated,” so “bright-X” and “X” share a centroid, but this is not universal.

Aimed at the problem that traditional histogram is sensitive to illumination changes in visual tracking, combined with the CN(Color Name) feature, we proposed a new feature(denotes CNH, Color Name Histogram) based on color name. Firstly, the method projected the original RGB image to CN space to obtain robust 11 feature layers. Then, we counted the each pixel numbers of feature layers. Finally, normalizing the amount of pixels in each layer. In addition, we adopted a feature adaptive fusion method to combine CNH and HOG(Histogram of Oriented Gradient). In order to prove validity of the proposed algorithm, we use Staple(Sum of Template And Pixel-wise Learners) algorithm frame to make a controlled trial. In contrast with the reference algorithms, the success of our algorithm increases by 1.5% and the precision increases by 1.7%. The results show that this method retains the advantages of traditional histogram which is insensitive to target deformation, but also enhances the robustness to illumination change.

Color lexis constitutes a complex semantic network of meanings, which sometimes makes it difficult to determine the characteristic feature of a given color name. This article analyzes the semantic ambiguity of color lexis in the Polish and Russian languages. Color ambiguity, as shown by the analysis, is related to the specificity of the color pattern, the lack of stability of the color pattern or its multicolour quality.

This paper aims to find an effective method for identifying colors of fabrics. We consider the following color naming models: Basic Color terms, ISCC-NBS Color system, Frery’s Color name, and Fractals Lab’s Ultimate color vocabulary. The appropriate color name is assigned for a fabric image by using seven types of minimum distance in RGB and HSV color models. Experiments are performed on S3 plain color fabric images. Results show that in general, the method of identifying fabric color name by using Frery’s color name model with the color similarity measure by Euclidean distance in RGB space gives the better result than the other methods. When we analyze in more details by classifying fabric images into three groups based on color saturation, we found that the effective methods for color naming are as follows: Frery’s color name model with the quadratic distance in HSV space works well for the group of low color saturation (96.23% accuracy). Frery’s color name model with Euclidean distance in RGB space works well for the group of medium color saturation (81.82% accuracy). And Fractals Lab’s Ultimate color vocabulary model with the weighted Euclidean distance in HSV space work good for the group of high color saturation (76.09% accuracy).

The color blue is problematic for inclusion in a symbology color set for use by operators with head-up and helmet-mounted displays (HUDs and HMDs), as well as with augmented reality (AR) displays. The distinguishing feature of these see-through displays is the optical combiner that presents color-coded symbology combined with the forward real-world scene. Unlike head-down displays that can use color fills, see-through displays are limited to lines and text. The presence of high-ambient daylight can desaturate symbol colors and make them difficult to recognize. This is especially true for blue symbols that become faint and colorless in moderate daylight as well as when mixed with green night-vision imagery. We propose the alternative color blue prime (blue')—a mix of 100% blue and 50% green that lies between blue and cyan—as an alternative to blue. The color blue’ is more resistant to daylight than blue, yet still retains the color name “blue.” Blue' lags other colors such as green and white in visibility, and its use needs to be moderated. We present experimental data to support the use of blue’ as a basic color code.

Designing a robust and accurate object tracker is important in many computer vision applications. The problem becomes more complicated when additional factors like changing appearance, illumination, and scale are introduced in the sequence. Recently, trackers that are based on the correlation filter method like Sum of Template and Pixel-wise Learners (STAPLE)1 have shown state-of-the-art short-term tracking performance. STAPLE consists two major modules: learning correlation filter on HOG features and representing color information using RGB histogram. In this paper, we propose an improved STAPLE (iSTAPLE) tracker by adding the Color Names (CN)2 to the correlation part of the tracker. CN complements HOG feature because using only HOG can lead to tracking failures in some cases where occlusion or deformation is present. As the color information could be a confusing factor and unreliable in tracking due to the rapid illumination changes, Bhattacharyya distance is used to measure the color similarity between the target and surrounding area to decide whether the color information is helpful. Since we use multiple feature cues to improve tracking performance, a robust approach to fuse multiple features is required. To fully utilize all features and optimize the tracking result, numerous weight combinations assigned to each feature are tested. We show through comprehensive experiments on the VOT Challenge 2016 dataset3 that iSTAPLE obtains a gain of 25% in tracking robustness.

Automatic colorization is the process of adding color to greyscale images. We condition this process on language, allowing end users to manipulate a colorized image by feeding in different captions. We present two different architectures for language-conditioned colorization, both of which produce more accurate and plausible colorizations than a language-agnostic version. Through this language-based framework, we can dramatically alter colorizations by manipulating descriptive color words in captions., Comment: 6 pages

Throughout history, litterateurs have always attached great importance to the use of color words in their works. Tang poetry and Song lyrics, the most brilliant parts of Chinese literature history, are perfect examples in using color words to create artistic conception and express the thoughts and feeling. In this paper, high-frequency color words that appear in “All-Tang Poetry” (The collection of ancient Chinese poetry in 618–907 A.D.) and “All-Song Poetry” (The collection of ancient Chinese lyrics in 960–1279 A.D.) are chosen as research objects, and the optimal algorithm is employed to study the collocation of the color words in Tang poetry and Song lyrics by comparing the effects of PMI and Word2vec extraction collocation. The detailed analysis of works of twenty poets in Tang and Song Dynasties are also carried out from both macro and micro perspectives, from the use of color words in “All-Tang Poetry” and “All-Song Poetry” to that specific works. As a result, most “ Open image in new window (white)” in Tang poetry expresses frustration or sadness. “ Open image in new window (red)”, the most popular word used in Song lyrics, expresses femininity, indicating that “ Open image in new window (Tang Poetry expresses aspiration and ambition, and Song lyrics express personal emotions between lovers)”.

Altres ajuts: CERCA Programme/Generalitat de Catalunya Color description is one of the fundamental problems of image understanding. One of the popular ways to represent colors is by means of color names. Most existing work on color names focuses on only the eleven basic color terms of the English language. This could be limiting the discriminative power of these representations, and representations based on more color names are expected to perform better. However, there exists no clear strategy to choose additional color names. We collect a dataset of 28 additional color names. To ensure that the resulting color representation has high discriminative power we propose a method to order the additional color names according to their complementary nature with the basic color names. This allows us to compute color name representations with high discriminative power of arbitrary length. In the experiments we show that these new color name descriptors outperform the existing color name descriptor on the taskof visual tracking, person re-identification and image classification.

This paper presents a set of methods for classifying the color attribute of objects when multiple images of the same objects are available. This problem is more complex than the single image estimation since varying environmental effects, such as, shadows or specularities from light sources, can result in poor accuracy. These depend primarily on the camera positions and the material type of the objects. Single image techniques focus on improving the discrimination of between colors, whereas in multi-view systems additional information is available but should be utilized wisely. To this end, we propose three methods to aggregate image pixel information in multi-view that boost the performance of color name classification. Moreover, we study the effect of shadows by employing automatic shadow detection and correction techniques on the color naming problem. We tested our proposals on a new multi-view color names dataset (M3DCN) which contain indoor and outdoor objects. The experimental evaluation shows that one out of the three presented aggregation methods is very efficient and it achieves the highest accuracy in term of classification results. Also, we experimentally show that addressing visual outliers like shadow in multi-view images improves the performance of the color attribute decision process.

Color blindness is commonly known as color vision deficiency, the vision defect which presented in our eye ball recognition of color is the major problem for the people who are having color blindness. In this work, we are proposing a hand held real-time automated guidance system to detect and label the color name for color blind persons that will help to identify and label color which they can't do using natural eyes. We designed this Real-time application using Matlab Simulink model. This Paper mainly focusing and developing the users interface it can detect the color and label the color name which is shown by using video glasses. This system mainly consists of computer, raspberry pi B, camera module, and video glasses.

In a text dating between 1259 and 1277, the Persian scholar al-Tusi presented a systematic arrangement of 26 color terms. We propose a reconstruction of all color terms from al-Tusi's scheme, in terms of preferred translation, mean CIEL*a*b* coordinates and digital representation. This reconstruction is based on a visual experiment with 30 subjects, who identified the Munsell chip best representing each color term. Persian words for which the meaning changed since the time of al-Tusi were substituted by direct translations. The results show considerable interobserver variability in the colors selected when identifying color terms. This relatively large variation was shown to be a characteristic for memory matching experiments in general. Several specific color terms for which the resulting color variation was particularly large are discussed in more detail, and possible explanations for these variations are proposed. The proposed reconstruction suggests that al-Tusi's list is largely consistent in modern colorimetric terms, although some large hue shifts are observed for color terms corresponding to green. We found no evidence for blue-green (“grue”) confusion. © 2015 Wiley Periodicals, Inc. Col Res Appl, 2015

We explore the following two cognitive questions regarding crosslinguistic variation in lexical semantic systems: Why are some linguistic categories-that is, the associations between a term and a portion of the semantic space-harder to learn than others? How does learning a language-specific set of lexical categories affect processing in that semantic domain? Using a computational word-learner, and the domain of color as a testbed, we investigate these questions by modeling both child acquisition of color terms and adult behavior on a non-verbal color discrimination task. A further goal is to test an approach to lexical semantic representation based on the principle that the more languages label any two situations with the same word, the more conceptually similar those two situations are. We compare such a crosslinguistically based semantic space to one based on perceptual similarity. Our computational model suggests a mechanistic explanation for the interplay between term frequency and the semantic closeness of learned categories in developmental error patterns for color terms. Our model also indicates how linguistic relativity effects could arise from an acquisition mechanism that yields language-specific topologies for the same semantic domain. Moreover, we find that the crosslinguistically inspired semantic space supports these results at least as well as-and in some aspects better than-the purely perceptual one, thus confirming our approach as a practical and principled method for lexical semantic representation in cognitive modeling.

The article is devoted to the research on the national specific features of color symbolism embodied in color words in the Russian and Chinese languages. In a language system, each color is characterized by cognitive, emotional and information functions for all the peoples of the world. However, in a particular culture each color serves as a special symbol, which can be associated by native speakers with similar or different emotions and reflect rich potential of national cultures. Color words, having symbolic properties, are not only important means to describe the world, but also they perform a cultural-linguistic function to convey emotions of people belonging to different nations. In the Russian and Chinese languages, there are many color words with a strong national flavor. Apart from evoking rich associations and expressing emotional experiences and feelings, they also form a distinct national color culture and have a great influence on human life of the two countries. The vivid contrast between the two languages helps to comprehend the semantic motivation of color words, shades of meanings of color symbolism and cultural identity.

Human experience of color results from a complex interplay of perceptual and linguistic systems. At the lowest level of perception, the human visual system transforms the visible light portion of the electromagnetic spectrum into a rich, continuous three-dimensional experience of color. Despite our ability to perceptually discriminate millions of different color shades, most languages categorize color into a number of discrete color categories. While the meanings of color words are constrained by perception, perception does not fully define them. Once color words are acquired, they may in turn influence our memory and processing speed for color, although it is unlikely that language influences the lowest levels of color perception. One approach to examining the relationship between perception and language in forming our experience of color is to study children as they acquire color language. Children produce color words in speech for many months before acquiring adult meanings for color words. Research in this area has focused on whether children’s difficulties stem from (a) an inability to identify color properties as a likely candidate for word meanings, or alternatively (b) inductive learning of language-specific color word boundaries. Lending plausibility to the first account, there is evidence that children more readily attend to object traits like shape, rather than color, as likely candidates for word meanings. However, recent evidence has found that children have meanings for some color words before they begin to produce them in speech, indicating that in fact, they may be able to successfully identify color as a candidate for word meaning early in the color word learning process. There is also evidence that prelinguistic infants, like adults, perceive color categorically. While these perceptual categories likely constrain the meanings that children consider, they cannot fully define color word meanings because languages vary in both the number and location of color word boundaries. Recent evidence suggests that the delay in color word acquisition primarily stems from an inductive process of refining these boundaries.

It has often been observed that color is a highly preferred attribute for use in distinguishing descriptions, that is, referring expressions produced with the purpose of identifying an object within a visual scene. However, most of these observations were based on visual displays containing only colors that were maximally different in hue and for which the language of experimentation possessed basic color terms. The experiments described in this paper investigate whether speakers' preference for color is reduced if the color of the target referent is similar to that of the distractors. Because colors that look similar are often also harder to distinguish linguistically, we also examine the impact of the codability of color values. As a third factor, we investigate the salience of available alternative attributes and its impact on the use of color. The results of our experiments show that, while speakers are indeed less likely to use color when the colors in a display are similar, this effect is mostly due to the difficulty in naming similar colors. Color use for color with a basic color term is affected only when the colors of target and distractors are very similar (yet still distinguishable). The salience of our alternative attribute size, manipulated by varying the difference in size between target and distractors, had no impact on the use of color.

Identifying objects by its color name is an essential capability for a service robot to understand and interact with the physical world and be of use in everyday life scenarios. The robot uses an image for the identification, but the image is strongly affected by lightning environment while human does not affected by the environment when they decide its color name. We present lightning environment estimation and perceptual color classification based on the estimation using hyperspectral data. Support Vector Machine (SVM) or Multiple Kernel Learning SVM is used for the estimation and classification. Originality of our paper is that the hyperspectral data is used for both lightning environment estimation and perceptual color classification of an object. Suppose that the lightning environment is given, the perceptual color of the object is classified in each environment. Additionally, perceptual color is estimated by using result of the lightning environment estimation. A novel dataset which consists of the hyperspectral image of 15 objects taken in 4 different lightning environments is constructed in our experiment. Experimental result of the hyperspectral data are compared by those of common color spaces, such as RGB and L*a*b. The estimation and classification with SVM and MKL SVM are compared in the experiments. From the experimental results, hyperspectral data enabled us to present a separated scheme which consists of the lightning environment estimation and perceptual color classification while previous color spaces could not.

Color serves as an important cue for many computer vision tasks. Nevertheless, obtaining accurate color description from images is non-trivial due to varying illumination conditions, view angles, and surface reflectance. This is especially true for the challenging problem of pedestrian description in public spaces. We made two contributions in this study: (1) We contribute a large-scale pedestrian color naming dataset with 14,213 hand-labeled images. (2) We address the problem of assigning consistent color name to regions of single object’s surface. We propose an end-to-end, pixel-to-pixel convolutional neural network (CNN) for pedestrian color naming. We demonstrate that our Pedestrian Color Naming CNN (PCN-CNN) is superior over existing approaches in providing consistent color names on real-world pedestrian images. In addition, we show the effectiveness of color descriptor extracted from PCN-CNN in complementing existing descriptors for the task of person re-identification. Moreover, we discuss a novel application to retrieve outfit matching and fashion (which could be difficult to be described by keywords) with just a user-provided color sketch.

This paper presents a person re-identification algorithm using HSV color histogram matching and sparse representation. To continuously track a person who is captured by multiple cameras without overlapping regions, the person re-identification is essential. The proposed algorithm uses color information to recognize a person in a low resolution image. The proposed person re-identification algorithm consists of two steps: i) HSV color histogram matching and gallery reduction and ii) sparse representation matching using color name descriptor. The proposed algorithm minimizes unnecessary calculations and increases accuracy by reducing the number of candidates using HSV color histogram matching. In addition, it is invariant to occlusion and illumination variation using the sparse representation-based re-identification approach. As a results, it can be applied to a multiple camera-based video surveillance system for public security in airports and stores.