Color Space Selection Research Articles

Exploring adversarial deep learning for fusion in multi-color channel skin detection applications

Deep learning, a robust framework for complex learning, outperforms previous machine learning approaches and finds widespread use. However, security vulnerabilities, especially in fusion in multi-color channel skin detection applications using adversarial machine learning (AML) and generative adversarial networks (GANs), lead to misclassifications. Researchers are actively exploring AML's and GANs' impact on misclassification, focusing on vulnerabilities in lighting conditions, skin-like patches in lesion segmentation, and insufficient data in facial emotion recognition. Yet, these areas only scratch the surface of potential AML vulnerabilities and GANs. To comprehensively address challenges, an in-depth investigation into AML and GANs components is crucial to uncover underlying reasons for misclassifying skin detection. This study addresses challenges of fusion in multi-color channel skin detection by creating a diverse dataset with 17M patches for enhanced feature fusion/training and meeting dataset criteria, investigating misclassifications using various deep learning models belonging to AML and GANs and color spaces (e.g., RGB, YCbCr, HSV, YUV), and exploring binary and multiclass scenarios. Notably, YCbCr outperformed RGB, achieving 98 % for binary skin classification, 84 % and 69 % for multiclass four and five-class scenarios. Binary classification for skin tones and their skin-like counterparts (e.g., black skin tone and black skin-like) yielded 97 %, 81 %, 60 %, and 51 % for black, brown, medium, and fair, respectively. Exploration of darker skin tones showed improved accuracy. Benchmarking with a CNN and RNN hybrid achieved 99 % accuracy, surpassing the initial 91 %, while SAE reached 97 %. The study explores implications of overlapping between skin and skin-tone recognition, offering insights for developing a generalized skin detector. The investigation demonstrates that improper color space selection can make lighting conditions exploitable in AML attacks and GANs, emphasizing the crucial role of color space choice in mitigating vulnerabilities.

Impact of Color Space and Color Resolution on Vehicle Recognition Models.

In this study, we analyze both linear and nonlinear color mappings by training on versions of a curated dataset collected in a controlled campus environment. We experiment with color space and color resolution to assess model performance in vehicle recognition tasks. Color encodings can be designed in principle to highlight certain vehicle characteristics or compensate for lighting differences when assessing potential matches to previously encountered objects. The dataset used in this work includes imagery gathered under diverse environmental conditions, including daytime and nighttime lighting. Experimental results inform expectations for possible improvements with automatic color space selection through feature learning. Moreover, we find there is only a gradual decrease in model performance with degraded color resolution, which suggests the need for simplified data collection and processing. By focusing on the most critical features, we could see improved model generalization and robustness, as the model becomes less prone to overfitting to noise or irrelevant details in the data. Such a reduction in resolution will lower computational complexity, leading to quicker training and inference times.

Exploring Image Representation and Color Spaces in Computer Vision: A Comprehensive Review

This paper presents a comprehensive review of image representation and color spaces in the domain ofcomputer vision. Image representation serves as the foundation of computer vision systems, encompassingtechniques such as pixel-based, vector-based, and feature-based representations. Color spaces provide astandardized framework for encoding color information in digital images, with popular models includingRGB, HSV, Lab, and CMYK. The paper explores fundamental concepts, comparative analysis, practicalapplications, and future directions in image representation and color spaces. Insights gained from the reviewhighlight the significance of these concepts in various computer vision applications, including objectrecognition, image segmentation, and image enhancement. Future research directions include addressingchallenges such as achieving color constancy and developing adaptive color space selection techniques. Byleveraging the findings from this review, researchers and practitioners can advance the state-of-the-art incomputer vision and develop more robust and effective systems for real-world applications.

RandStainNA++: Enhance Random Stain Augmentation and Normalization Through Foreground and Background Differentiation.

The wide prevalence of staining variations in digital pathology presents a significant obstacle, often undermining the effectiveness of diagnosis and analysis. The current strategies to counteract this issue primarily revolve around Stain Normalization (SN) and Stain Augmentation (SA). Nonetheless, these methodologies come with inherent limitations. They struggle to adapt to the vast array of staining styles, tend to presuppose linear associations between color spaces, and often lead to unrealistic color transformations. In response to these challenges, we introduce RandStainNA++, a novel method seamlessly integrating SN and SA. This method exploits the versatility of random SN and SA within randomly selected color spaces, effectively managing variations for the foreground and background independently. By refining the transformations of staining styles for the foreground and background within a realistic scope, this strategy promotes the generation of more practical staining transformations during the training phase. Further enhancing our approach, we propose a unique self-distillation method. This technique incorporates prior knowledge of stain variation, substantially augmenting the generalization capability of the network. The striking results yield that, compared to conventional classification models, our method boosts performance by a significant margin of 16-25%. Furthermore, when juxtaposed with baseline segmentation models, the Dice score registers an increase of 0.06.

Critical importance of RGB color space specificity for colorimetric bio/chemical sensing: A comprehensive study

The use of the RGB color model in colorimetric chemical sensing via imaging techniques is widely prevalent in the literature. However, the lack of specificity in the selection of RGB color space during capture and analysis presents a significant challenge in creating standardised methods for this field and possible discrepancies. In this study, we conducted a comprehensive comparison and contrast of a total of 68 RGB color spaces to evaluate their respective impacts on colorimetric bio/chemical sensing. We explore the impact of dynamic range, sensitivity, and limit of detection, and show that the lack of specificity in RGB color space selection can significantly impact colorimetric chemical sensing by 42–77%. We also explore the impact of underlying RGB comparisons and demonstrate a further 18.3% discrepancy between RGB color spaces. By emphasising the importance of proper RGB color space selection and handling, our findings contribute to a better understanding of this critical area and present valuable opportunities for future research. We further provide valuable insights for creating standardised methods in this field, which can be utilised to avoid discrepancies and ensure accurate and reliable analysis in colorimetric bio/chemical sensing.

Design of Machine Learning Solutions to Post-Harvest Classification of Vegetal Species

This paper presents a machine learning approach to automatically classifying post-harvest vegetal species. Color images of vegetal species were applied to convolutional neural networks (CNNs) and support vector machine (SVM) classifiers. We focused on okra as the target vegetal species and classified it into two quality types. However, our approach could also be applied to other species. The machine learning solution consists of several components, and each design process and its combinations are essential for classification quality. Therefore, we carefully investigated their effects on classification accuracy. Through our experimental evaluation, we confirmed the following: (1) in color space selection, HLG (hue, lightness, and green) and HSL (hue, saturation, and lightness) are essential for vegetal species; (2) suitable preprocessing techniques are required owing to the complexity of the data and noise load; and (3) the diversity extension of learning image data by mixing different datasets obtained under different conditions is quite effective in reducing the overfitting possibility. The results of this study will assist AI practitioners in the design and development of post-harvest classifications based on machine learning.

Facial emotion recognition using deep learning detector and classifier

Numerous research works have been put forward over the years to advance the field of facial expression recognition which until today, is still considered a challenging task. The selection of image color space and the use of facial alignment as preprocessing steps may collectively pose a significant impact on the accuracy and computational cost of facial emotion recognition, which is crucial to optimize the speed-accuracy trade-off. This paper proposed a deep learning-based facial emotion recognition pipeline that can be used to predict the emotion of detected face regions in video sequences. Five well-known state-of-the-art convolutional neural network architectures are used for training the emotion classifier to identify the network architecture which gives the best speed-accuracy trade-off. Two distinct facial emotion training datasets are prepared to investigate the effect of image color space and facial alignment on the performance of facial emotion recognition. Experimental results show that training a facial expression recognition model with grayscale-aligned facial images is preferable as it offers better recognition rates with lower detection latency. The lightweight MobileNet_v1 is identified as the best-performing model with WM=0.75 and RM=160 as its hyper-parameters, achieving an overall accuracy of 86.42% on the testing video dataset.

Retinal image preprocessing techniques: Acquisition and cleaning perspective

AbstractImage preprocessing is a method to transform raw image data into clean image data. The objective of preprocessing is to improve the image data by suppressing undesired distortions. Enhancement of some image features which are relevant for further processing of image and analysis task is also done in preprocessing. Screening and diagnosis of various eye diseases like diabetic retinopathy, Choroidal Neovascularization(CNV), DRUSEN, etc. are possible using digital retinal images. This paper aims to provide a better understanding and knowledge of the computer algorithms used for retinal image preprocessing. In this paper, various image preprocessing techniques are incorporated such as color correction, color space selection, noise reduction, and contrast enhancement on retinal images. Retinal blood vessels are better seen in Green color space instead of Red or Blue color space. Noise reduction through Block matching and 3D(BM3D) techniques show a significant result as compared to Total Variation Filter (TVF) and Bilateral Filter (BLF). Contrast enhancement through Contrast Limited Adaptive Histogram Equalization (CLAHE) outperforms Global Equalization (GE) or Adaptive Histogram Equalization (AHE). Evaluation parameters such as Mean square error, Peak Signal Noise ratio, Structured similarity index measures, and Normalized root mean square error values for BM3D noise filtering are 0.0029, 25.3370, 0.6839 and 0.0998 respectively which shows that BM3D outperforms the others.

Significance of Color Spaces and Their Selection for Image Processing: A Survey

: The human visual system is encompassed with three components that help to produce a color sensation. The color display devices use this concept and create their full range of colors by incorporating all three primary color components. Any displayable color can be created with the help of these primaries. This paper comprises the concept of color space, which helps to understand a particular device’s color proficiencies. Since colors are a more robust descriptor, color spaces are considered to play a significant role in representing an image appropriately. The performance of many algorithms depends on the selection of an appropriate color space. There are many kinds of color spaces that can be used in real-time applications. These are RGB, nRGB, HSV, TSL, YCbCr, YUV, YES, CIE-XYZ, CMYK, etc. In this paper, color spaces and their vast classification are described in brief. Apart from these, the characteristics, applicability, limitations, mathematical conversion, and other essential factors are also explained for each color space.

Classification and detection using hidden Markov model‐support vector machine algorithm based on optimal colour space selection for blood images

Patients with cerebral haemorrhages need to drain haematomas. Fresh blood may appear during the haematoma drainage process, so this needs to be observed and detected in real time. To solve this problem, this paper studies images produced during the haematoma drainage process. A blood image feature selection recognition and classification framework is designed. First, aiming at the characteristics of the small colour differences in blood images, the general RGB colour space feature is not obvious. This study proposes an optimal colour channel selection method. By extracting the colour information from the images, it is recombined into a 3 × 3 matrix. The normalised 4-neighbourhood contrast and variance are calculated for quantitative comparison. The optimised colour channel is selected to overcome the problem of weak features caused by a single colour space. After that, the effective region in the image is intercepted, and the best colour channel of the image in the region is transformed. The first, second and third moments of the three best colour channels are extracted to form a nine-dimensional eigenvector. K-means clustering is used to obtain the image eigenvector, outliers are removed, and the results are then transferred to the hidden Markov model (HMM) and support vector machine (SVM) for classification. After selecting the best color channel, the classification accuracy of HMM-SVM is greatly improved. Compared with other classification algorithms, the proposed method offers great advantages. Experiments show that the recognition accuracy of this method reaches 98.9%.

CU-Net: A New Improved Multi-Input Color U-Net Model for Skin Lesion Semantic Segmentation

Melanoma is considered one of the most dangerous skin cancer diseases that threaten human health and life. Early diagnosis of melanoma is a big challenge, especially with the presence of color variations across similar lesion types. Automatic skin lesion segmentation is an essential step to build a successful skin disease classification system. Recent deep learning architectures significantly improve the skin lesion segmentation results. Especially, U-Net deep convolutional neural network (CNN) is considered one of the state-of-the-art models with promising performance. Most deep CNNs and particularly U-Net model utilize a single input RGB color image for skin lesion semantic segmentation. However, RGB color space is not usually the best choice to represent the invariant characteristics of skin lesion chromatic information. The selection of the optimal color space significantly affects the performance of segmentation results. In this paper, three novel variants of U-Net model with single, dual, and triple inputs, namely, Single Input Color U-Net (SICU-Net), Dual Input Color U-Net (DICU-Net) and Triple Input Color U-Net (TICU-Net) are proposed. The structure of SICU-Net, DICU-Net, and TICU-Net contains single, dual, and triple encoder sub-networks connected with only a single decoder path. Each encoder sub-network is fed with different color space of the input image. A channel-wise attention module is utilized to fuse the contribution of the learned feature maps from each encoder sub-network which is fed to the decoder sub-network to generate segmented image map. Moreover, a composite loss function is designed to improve the performance of the proposed CU-Net models. Three public benchmark datasets, namely, International Skin Imaging Collaboration (ISIC 2017, ISIC 2018) and PH2 datasets, are utilized to evaluate the performance of the proposed models. Experimental results reveal that the proposed models significantly improve the performance of the original U-Net model and achieve comparable performance with other state-of-the-art methods.

A method for analysing the color rendering of digital cameras

The paper proposes a method for studying the color rendition of digital cameras. This parameter is conventionally assessed visually by color targets in photography and design. The proposed method compares the chromaticity values of standard test objects with their measured values on a digital camera using special setup features. The setup includes a light source, a collimator for uniform illumination of a reflecting screen with a test object, and a research object, i.e., a digital camera. The camera is positioned at a 45 degree angle to the screen. This setup follows the recommendations of the Commission Internationale de l’Eclairage (CIE) for color measurements. To measure the colorimetric characteristics of the samples, the authors adopted the 0/45 illumination / observation scheme. The test object is illuminated with a light beam, the axis of which makes an angle not exceeding 10° from the normal to the sample surface. The sample is observed at an angle of 45° ± 5° from to the normal. The angle between the axis of the illuminating beam and any of its rays does not exceed 5°. A type A source (color temperature 2856 K) acts as an illuminator. Standardized reference sets of colored optical glasses (light filters) characterized by known chromaticity coordinates were chosen as test objects. To evaluate the results, specialized software has been developed that allows one to select individual pixels, calculate their brightness in order to find the chromaticity coordinates and compare the results with reference values. The method was tested using the Canon EOS 60D digital camera. At the time of measurement, digital filters for correction, anti-aliasing, sharpening, as well as color adjustment were turned off in the camera. The paper presents colorimetric measurements using 58 colored optical glasses. The averaged values fall into seven groups. The selected color space for measurements is sRGB. The measurement results proved the possibility of using the proposed technique for analysing and choosing appropriate digital recording devices for colorimetric measurements in such areas as medicine, chemistry and food industry.

CODEX weak lensing mass catalogue and implications on the mass–richness relation

ABSTRACT The COnstrain Dark Energy with X-ray clusters (CODEX) sample contains the largest flux limited sample of X-ray clusters at 0.35 < z < 0.65. It was selected from ROSAT data in the 10 000 square degrees of overlap with BOSS, mapping a total number of 2770 high-z galaxy clusters. We present here the full results of the CFHT CODEX programme on cluster mass measurement, including a reanalysis of CFHTLS Wide data, with 25 individual lensing-constrained cluster masses. We employ lensfit shape measurement and perform a conservative colour–space selection and weighting of background galaxies. Using the combination of shape noise and an analytic covariance for intrinsic variations of cluster profiles at fixed mass due to large-scale structure, miscentring, and variations in concentration and ellipticity, we determine the likelihood of the observed shear signal as a function of true mass for each cluster. We combine 25 individual cluster mass likelihoods in a Bayesian hierarchical scheme with the inclusion of optical and X-ray selection functions to derive constraints on the slope α, normalization β, and scatter σln λ|μ of our richness–mass scaling relation model in log-space: ${\langle {\rm In}\,\, \lambda\!\!\mid\!\!\mu\rangle = \alpha\mu + \beta,} $ with μ = ln (M200c/Mpiv), and Mpiv = 1014.81M⊙. We find a slope $\alpha = 0.49^{+0.20}_{-0.15}$, normalization $\exp (\beta) = 84.0^{+9.2}_{-14.8}$, and $\sigma _{\ln \lambda | \mu } = 0.17^{+0.13}_{-0.09}$ using CFHT richness estimates. In comparison to other weak lensing richness–mass relations, we find the normalization of the richness statistically agreeing with the normalization of other scaling relations from a broad redshift range (0.0 < z < 0.65) and with different cluster selection (X-ray, Sunyaev–Zeldovich, and optical).

Color space and color channel selection on image segmentation of food images

Image segmentation is a predefined process of image processing to determine a specific object. One of the problems in food recognition and food estimation is the lack of quality of the result of image segmentation. This paper presents a comparative study of different color space and color channel selection in image segmentation of food images. Based on previous research regarding image segmentation used in food leftover estimation, this paper proposed a different approach to selecting color space and color channel based on the score of Intersection Over Union (IOU) and Dice from the whole dataset. The color transformation is required, and five color spaces were used: CIELAB, HSV, YUV, YCbCr, and HLS. The result shows that A in LAB and H in HLS are better to produce segmentation than other color channels, with the Dice score of both is 5 (the highest score). It concludes that this color channel selection is applicable to be embedded in the Automatic Food Leftover Estimation (AFLE) algorithm.

Enhanced Unsupervised Change Detector for Industrial Surveillance Systems

Detecting an intruder that is trespassing a prohibited area is a critical task of intelligent surveillance systems. This task requires a change detector to segment an intruder (foreground object) from the background. The task suffers the inherent drawbacks of change detectors due to the dual-camera sensor (color/IR), illumination changes, night time, static, and camouflaged foreground objects. This article proposes an enhanced unsupervised change detector to compensate for the aforementioned challenges for industrial sterile zone monitoring. The camera switch detection based on skewness patterns detects a switch between the dual-camera sensors (color/IR). The optimal color space selection based on the mean squared error will select tolerant color space (RGB/YCbCr) to illumination changes for modeling the background. In addition, the IR camera frames are contrast-enhanced to tackle the camouflaged intruders during the night. The incoming frames are split into respective channels before modeling the background. The background is modeled by Gaussian mixture models. The adaptive background model update scheme is proposed to tackle the various challenges posed by environment and object such as a static foreground object. The comparison is performed on three databases with top-ranked unsupervised change detection algorithms.

Strip-pair comparison method for building threshold color-difference model: theoretical model validation.

This paper presents a method for developing color-difference models near a threshold, based on the serial exploration method described by Torgerson [Theory and Methods of Scaling; Wiley & Sons (1958); Chap. 7], involving the construction of color-control strips of patches arranged in arrays of 2 × n, where n is the number of pairs in the strip. The patches in the lower row should be calorimetrically identical, while the color of the patches in the upper row should vary progressively in constant steps of CIELAB color difference along selected color space vector directions. Prospective observers are instructed to indicate the patch pair number for which they begin to perceive a slight color difference between corresponding patches. The frequency data obtained from the observers was used to build a threshold color-difference model. The intention was to validate the method with theoretical data to determine the effect of the precision with which the strips are constructed, on the accuracy of the estimated parameters. Theoretical frequency data was generated using the CIE94 color difference formula, whose associated color discrimination ellipsoid parameters are very easy to determine, associated with a hypothetical logistic psychometric curve for different color centers. The proposed method allows to determine color discrimination parameters with a precision nearby 4% and an accuracy of 3% with respect to the simulated theoretical parameters, for color samples generated with a standard deviation of ΔEab∗=0.2 of the superimposed error around the ideal color difference of pairs of patches.

Recognition of Mobile Robot Navigation Path Based on K-Means Algorithm

With the rapid development of computer technology and electronics industry, computer processing capability and image processing technology have been greatly improved, making robots based on computer processing and image processing have entered a new development in the field of navigation path recognition research. As an indispensable carrier for intelligent manufacturing and industrial development, robots are expanding their applications. The key to the successful execution of the mobile robot is to move according to the planned path and to avoid obstacles autonomously. These two points depend on the validity and accuracy of navigation path identification. At present, research on mobile robot navigation path recognition mainly uses visual navigation as the main method, which uses visual sensors to simulate human eye functions, obtains relevant information from external environment images, and processes them to realize related functions that the system needs to complete. The two major problems in visual navigation are poor recognition ability and insufficient ability to resist light source interference. The main purpose of this paper is to improve the recognition ability of mobile robot navigation path and the ability to resist light source interference. It mainly uses the [Formula: see text]-means algorithm for visual navigation research. By simulating the acquired image and the selected color space, the results show that the average time taken to complete a path identification method is 152[Formula: see text]ms. Under different illumination environments, the information extraction rate of mobile robot navigation path can reach 90%, and the effect of strong light on navigation path recognition is effectively reduced under strong illumination environment. The results show that the recognition of the visual navigation path of a mobile robot using the [Formula: see text]-means algorithm is more precise than the conventional method, and it takes less time to better meet the timeliness requirements of mobile robots.

Detection of hemorrhage in retinal images using linear classifiers and iterative thresholding approaches based on firefly and particle swarm optimization algorithms

We propose a novel iterative thresholding approach based on firefly and particle swarm optimization to be used for the detection of hemorrhages, one of the signs of diabetic retinopathy disease. This approach consists of the enhancement of the image using basic preprocessing methods, the segmentation of vessels with the help of Gabor and Top-hat transformation for the removal of the vessels from the image, the determination of the number of regions with hemorrhages and pixel counts in these regions using firefly algorithm (FFA) and particle swarm optimization algorithm (PSOA)-based iterative thresholding, and the detection of hemorrhages with the help of a support vector machine (SVM) and linear regression (LR)-based classifier. In the preprocessing step, color space selection, brightness and contrast adjustment, and adaptive histogram equalization are applied to enhance retinal images, respectively. In the step of segmentation, blood vessels are detected by using Gabor and Top-hat transformations and are removed from the image to avoid confusion with hemorrhagic regions in the retinal image. In the iterative thresholding step, the number of hemorrhagic regions and pixel counts in these regions are determined by using an iterative thresholding approach that generates different thresholding values with the FFA/PSOA. In the classification step, the hemorrhagic regions and pixel counts obtained by the iterative thresholding are used as inputs in the LR/SVM-based classifier. PSOA-based iterative thresholding and the SVM classifier achieved 96.7 % sensitivity, 91.4 % specificity, and 94.1 % accuracy for hemorrhage detection. Finally, the experiments show that the correct classification rates and time performances of the PSOA-based iterative thresholding algorithm are better than those of the FFA in hemorrhage detection. In addition, the proposed approach can be used as a diagnostic decision support system for detecting hemorrhages with high success rate.

A unified framework for exploiting color coefficients for salient object detection

Abstract With many applications in high level scene understanding, salient object detection is an important objective. In addition, this topic has received significant attention recently. This work proposes techniques that exploit the role of color spaces for addressing two important challenges in relation to salient object detection. The autonomous identification of a color space by which to carry out the processing has always been of prime importance to most image analysis tasks and this is equally so in relation to saliency detection. To address this challenge, a new adaptive color space selection method is proposed, here, which autonomously identifies the color space that is locally the most appropriate for saliency detection on an image by image basis. The color channels within this identified local color space are aggregated using joint l2, 1-norm minimization in order to determine feature importance and also to achieve feature selection leaning. A process relevant to saliency detection is multi-modality feature fusion, in which multiple features and color spaces are used in combination to capture the saliency aspects of an image/region. To implement this second process, a new technique for the region based optimal combination of feature modalities is introduced. The results of the rigorous experimental evaluations demonstrate the effectiveness of the adaptive color space selection and also for the region based optimal combination methods in comparison to 13 other state-of-the-art saliency methods and all in relation to three benchmark datasets. The efficacy of the proposed color space selection and region based combination methods is further validated by examining their ability to select appropriate color spaces for, and successfully aggregate the results of, up to five benchmark methods.

Perceptual Color Space Representations in the Oculomotor System Are Modulated by Surround Suppression and Biased Selection.

The oculomotor system utilizes color extensively for planning saccades. Therefore, we examined how the oculomotor system actually encodes color and several factors that modulate these representations: attention-based surround suppression and inherent biases in selecting and encoding color categories. We measured saccade trajectories while human participants performed a memory-guided saccade task with color targets and distractors and examined whether oculomotor target selection processing was functionally related to the CIE (x,y) color space distances between color stimuli and whether there were hierarchical differences between color categories in the strength and speed of encoding potential saccade goals. We observed that saccade planning was modulated by the CIE (x,y) distances between stimuli thus demonstrating that color is encoded in perceptual color space by the oculomotor system. Furthermore, these representations were modulated by (1) cueing attention to a particular color thereby eliciting surround suppression in oculomotor color space and (2) inherent selection and encoding biases based on color category independent of cueing and perceptual discriminability. Since surround suppression emerges from recurrent feedback attenuation of sensory projections, observing oculomotor surround suppression suggested that oculomotor encoding of behavioral relevance results from integrating sensory and cognitive signals that are pre-attenuated based on task demands and that the oculomotor system therefore does not functionally contribute to this process. Second, although perceptual discriminability did partially account for oculomotor processing differences between color categories, we also observed preferential processing of the red color category across various behavioral metrics. This is consistent with numerous previous studies and could not be simply explained by perceptual discriminability. Since we utilized a memory-guided saccade task, this indicates that the biased processing of the red color category does not rely on sustained sensory input and must therefore involve cortical areas associated with the highest levels of visual processing involved in visual working memory.