Color Space Conversion Research Articles

DEEP LEARNING FOR IMAGE CLASSIFICATION IN DIFFERENT COLOR SYSTEMS

Image classification is one of the most fundamental applications in the field of computer vision, which continues to arouse enormous interest. People can recognize a large number of objects in images with little effort, even though a number of their characteristics can change. Objects can be recognized even when their detection is partially difficult. At the same time, algorithmic description of the recognition problem for computer implementation is still an urgent problem. Existing methods for solving this problem are effective only for certain cases (for example, for geometric objects, human faces, road signs, printed or handwritten symbols) and only under certain conditions. A model designed to identify and classify objects must be able to identify their location, as well as distinguish between various features of objects, such as edges, corners, color differences, etc. Deep convolutional neural networks have demonstrated the best performance when working with images, which sometimes exceeds the capabilities of human vision. However, even with this significant improvement, there are still some issues with overfitting and vanishing gradient. To solve them, some well-known methods are used: data augmentation, batch normalization and dropout; modern classification models do not perform color space conversion of original RGB images. The study of the use of different color spaces in the task of image classification is one of the topical problems related to deep learning, and it determines the relevance of this study. If this problem is solved, this will improve performance of the models used.

Extraction of urban buildings using color image processing and morphological operations.

In many geospatial applications, automated detection of buildings has become a key concern in recent years. Determination of building locations provides great benefits for numerous geospatial applications such as urban planning, disaster management, infrastructure planning, environmental monitoring. The study aims to present a practical technique for extracting the buildings from high-resolution satellite images using color image segmentation and binary morphological image processing. The proposed method is implemented on satellite images of 4 different selected study areas of the city of Batikent, Ankara. According to experiments conducted on the study areas, overall accuracy, sensitivity, and F1 values were computed to be on average, respectively. After applying morphological operations, the same metrics are calculated . The results show that the determination of urban buildings can be done more successfully with the suitable combination of morphological operations using rectangular structuring element. Keywords: Building Extraction; Colour Image Processing;Colour space conversion; Image Morphology; Remote Sensing

Colour space conversion model from CMYK to CIELab based on CS‐WNN

AbstractIn the process of colour space conversion from CMYK to CIELab, colour difference will be caused, which has a negative impact on the quality of digital printing products. In this article, an improved wavelet neural network (WNN) model optimised by cuckoo search (CS) algorithm is proposed to reduce the colour difference. Initially, the colour space conversion model based on WNN is established. The CS algorithm is used to optimise the initial weights and parameters of dilation and translation in the WNN model. Then, 1296 samples are made to train the CS‐WNN model. Finally, 100 test samples are input into the trained network to obtain the corresponding L, a and b values of CIELab. The experimental results show that the average conversion colour difference () of the proposed model is 3.469. The conversion accuracy and stability of the proposed model are better than the traditional neural network.

Developing a Machine Vision System Equipped with UV Light to Predict Fish Freshness Based on Fish-Surface Color

This study assessed the feasibility of developing a machine vision system equipped with ultraviolet (UV) light, using changes in fish-surface color to predict aerobic plate count (APC, a standard freshness indicator) during storage. The APC values were tested and images of the fish surface were taken when fish were stored at room temperature. Then, images’ color-space conversion among RGB, HSV, and L*a*b* color spaces was carried out and analyzed. The results revealed that a* and b* values from the UV-light image decreased linearly during storage. A further regression analysis of these two parameters with APC value demonstrated a good exponential relationship between the a* value and the APC value (R2 = 0.97), followed by the b* (R2 = 0.85). Therefore, our results suggest that the change in color of the fish surface under UV light can be used to assess fish freshness during storage.

Study on Detection Image Processing Method of Offshore Cage

The paper aims to test the integrity of offshore cages. In this paper, offshore robots are used to collect video images. Through preprocessing methods such as color space conversion and Retinex algorithm of fusion-oriented filtering, the color distortion of the image is improved and the edge information of the image is maintained, so that the collected cage image is better recognized and calculated by the computer. And the paper uses the Canny algorithm to accurately locate the edge of the image, so that the boundary information can be reflected in more detail. In this paper, the adaptive threshold segmentation algorithm combined with mathematical morphology processing method is used to segment the offshore cage image, and extract the skeleton of the decomposed image to realize the detection of the integrity of the offshore net. Finally, a ROV equipped with a camera was used to verify the proposed method in a wave-flow water tank, which confirmed the effectiveness of the method and optimized the integrity recognition and image processing of the cage.

Research on Sugarcane Seed-Bud Location Based on Anisotropic Scaling Transformation

Highlights A method of locating sugarcane seed bud based on anisotropy transformation is proposed.Using computer binocular vision technology, the location of the sugarcane seed bud was determined by edge feature matching of the sugarcane seed bud.There are few methods to study the automatic location of sugarcane seed buds, and our research provides a new research idea.Abstract. Sugarcane is a major economic crop in China, but the degree of mechanization in sugarcane cultivation is low. To improve the economic benefit of sugarcane planting, promoting the use of mechanization in sugarcane planting is necessary. Currently, the sugarcane planted using mechanization has a low survival rate and the mechanization efficiency is low because the existing sugarcane precutting machine fails to address the problem of damaging seed buds. This study proposed a sugarcane bud localization method based on computer binocular vision technology. The sugarcane stem segment positions can be determined by the grayscale horizontal projection after preprocessing the sugarcane image based on color and grayscale features. Then, the bud area can be intercepted according to the positional relationship between the seed bud and the stem node, and the planar position of the seed bud will be determined by using the color space conversion and the gray vertical projection. Finally, the anisotropic scaling transformation is used to match the seed-bud area and restore the spatial coordinates of the seed bud, and the spatial position of the seed bud can be determined. The image pyramid acceleration matching process is adopted, which can make the method more suitable for real-time applications. The experimental results show that the accuracy of seed-bud matching based on the anisotropic scaling transformation is 98%, which provides a basis for research on the anti-injury germ system in the automatic planting process of sugarcane. Keywords: Anisotropic scaling, Binocular vision, Image pyramid, Mechanization planting of sugarcane, Seed bud location.

A Novel Feature Extraction for Complementing Authentication in Hand-based Biometric

With an increasing usage of hand-based biometrics in authentication system, there is a need to evolve up with more potential security owing to increasing evolution of threats. The security of the hand authentication system completely depends upon uniqueness and distinct selection of features from hand image which has the properties of robustness, fault tolerance, and simpler implication. Review of existing feature extraction literatures shows more inclination towards sophisticated process as well as it also suffers from various other limitation. Therefore, this manuscript resolves this limitation by presenting a novel model of feature extraction which is carried out in more progressive form and less iterative form, unlike existing approaches. The proposed system achieves its research goal by introducing simplified feature extraction operation via storage, blurring, color space conversion, binary image conversion, the modelling aspect of the study emphasizes on image enhancement along with fuzzification for yielding more efficient result. An experimental study has been carried out using Python considering hand-biometric dataset in order to carry analysis where the outcome shows significant supportability over any palmprint recognition system. The study outcome is compared with most standard implementation of feature extraction to find that proposed system offer better accuracy performance in contrast with existing system.

Radial Basis Function Neural Network for the Evaluation of Image Color Quality Shown on Liquid Crystal Displays

The color quality of an image shown on a liquid crystal display (LCD) can be measured with a spectroradiometer; however, this instrument is expensive, work under controlled illumination conditions with an artificial source of light, and measurements take a long time. A spectroradiometer returns measurements of wavelength or CIE color space. A low-cost and fast alternative consists of using a digital camera that outputs RGB measurements. Unfortunately, comparisons between measurements obtained with both instruments cannot be performed; hence, conversion equations must be used. The main problem is that equations do not consider the effects caused by the camera lens, sensor variations, and configurable parameters such as gain and the exposure time. This paper proposes the architecture of a radial basis function neural network (RBFNN) to measure the image color quality displayed by an LCD using a digital camera. The RGB values acquired with a camera are used as inputs to the RBFNN. The output predicted the luminance and chromaticity components in the CIExyY color space and included the corrections to the lens and camera parameters. First, the RBFNN topology is explained, including the calculation of the number of neurons in the hidden layer, and the definition of the dispersion centers and their associated spread. Next, the experiments related to RGB color space reconstruction and conversion from RGB to CIE are presented. The proposed approach was tested on a real automotive scenario. The results obtained were similar to those measured with the spectroradiometer with an accuracy of 93.3%. Moreover, the results remained within limits established by the six-sigma methodology.

Machine vision on the positioning accuracy evaluation of poultry viscera in the automatic evisceration robot system

ABSTRACT In the poultry slaughtering, accurate viscera positioning is essential to reduce the damage of internal organs. The introduction of machine vision technique can help to locate the viscera and can provide new direction for the poultry evisceration. After a midline abdominal incision of the poultry, the internal organs are taken out from the poultry that placed on the conveyor line by the parallel robot. Based on machine vision, the recognition accuracy of opened poultry viscera directly affects the level of visceral damage and residue caused by gripping manipulator. However, visceral positioning is often influenced by different noise in the abdominal cavity, such as mucous membranes and blood stains. Thus, the image segmentation of poultry viscera is a complex process, and it is challenging to remove the noise. In general, existing image segmentation methods can hardly segment visceral regions well. To strengthen the anti-noise ability, we proposed an improved region-based active contour method with the level-set formulation. This method combined with several operations of color space conversion and top-low-hat transformation, which could extract the viscera contour and precisely removed the noise. The results showed that recognition accuracy of the heart-liver area and fat area in the viscera are 98.98% and 99.75%, respectively, while the overall viscera for poultry was 98.96%. The results of this experiment suggested that the proposed image segmentation algorithm could achieve the required accuracy for poultry viscera detection. Thus, the proposed visceral contour recognition method could be applied in poultry processing, providing critical information to guide the robot for automated evisceration.

EVALUATING THE PERFORMANCE OF A SEMI-AUTOMATIC APPLE FRUIT DETECTION IN A HIGH-DENSITY ORCHARD SYSTEM USING LOW-COST DIGITAL RGB IMAGING SENSOR

Abstract: This study investigates the potential use of close-range and low-cost terrestrial RGB imaging sensor for fruit detection in a high-density apple orchard of Fuji Suprema apple fruits (Malus domestica Borkh). The study area is a typical orchard located in a small holder farm in Santa Catarina’s Southern plateau (Brazil). Small holder farms in that state are responsible for more than 50% of Brazil’s apple fruit production. Traditional digital image processing approaches such as RGB color space conversion (e.g., rgb, HSV, CIE L*a*b*, OHTA[I 1 , I 2 , I 3 ]) were applied over several terrestrial RGB images to highlight information presented in the original dataset. Band combinations (e.g., rgb-r, HSV-h, Lab-a, I” 2 , I” 3 ) were also generated as additional parameters (C1, C2 and C3) for the fruit detection. After, optimal image binarization and segmentation, parameters were chosen to detect the fruits efficiently and the results were compared to both visual and in-situ fruit counting. Results show that some bands and combinations allowed hits above 75%, of which the following variables stood out as good predictors: rgb-r, Lab-a, I” 2 , I” 3 , and the combinations C2 and C3. The best band combination resulted from the use of Lab-a band and have identical results of commission, omission, and accuracy, being 5%, 25% and 75%, respectively. Fruit detection rate for Lab-a showed a 0.73 coefficient of determination (R2), and fruit recognition accuracy rate showed 0.96 R2. The proposed approach provides results with great applicability for small holder farms and may support local harvest prediction.

基于改进极限学习机的数码印花颜色空间转换方法

基于改进极限学习机的数码印花颜色空间转换方法

DetexNet: Accurately Diagnosing Frequent and Challenging Pediatric Malignant Tumors.

The most frequent extracranial solid tumors of childhood, named peripheral neuroblastic tumors (pNTs), are very challenging to diagnose due to their diversified categories and varying forms. Auxiliary diagnosis methods of such pediatric malignant cancers are highly needed to provide pathologists assistance and reduce the risk of misdiagnosis before treatments. In this paper, inspired by the particularity of microscopic pathology images, we integrate neural networks with the texture energy measure (TEM) and propose a novel network architecture named DetexNet (deep texture network). This method enforces the low-level representation pattern clearer via embedding the expert knowledge as prior, so that the network can seize the key information of a relatively small pathological dataset more smoothly. By applying and finetuning TEM filters in the bottom layer of a network, we greatly improve the performance of the baseline. We further pre-train the model on unlabeled data with an auto-encoder architecture and implement a color space conversion on input images. Two kinds of experiments under different assumptions in the condition of limited training data are performed, and in both of them, the proposed method achieves the best performance compared with other state-of-the-art models and doctor diagnosis.

Internet of Things and Cloud Enabled Hybrid Feature Extraction with Adaptive Neuro Fuzzy Inference System for Diabetic Retinopathy Diagnosis

At present times, the diabetic retinopathy (DR) become high and it is required to design an Internet of Things (IoT) enabled DR diagnosis tool to assist the diagnosis process of remote patients. This study designs and develops IoT and cloud computing based Hybrid Feature Extraction (HFE) with Adaptive Neuro Fuzzy Inference System (ANFIS) for DR detection and classification model, abbreviated as HFE-ANFIS model. The proposed model initially captures the retinal fundus image of the patient using the IoT enabled head mounted camera and transmit the images to the cloud server, which executes the diagnosis process. The image preprocessing takes place using three stages namely color space conversion, filtering, and contrast enhancement. Next, segmentation process takes place using fuzzy c-means (FCM) model to identify the diseased portions in the fundus image. Then, HFE based feature extraction and ANFIS based classification processes are carried out to grade the different levels of DR. The performance validation of the HFE-ANFIS model takes place against MESSIDOR dataset and the results are investigated under different dimensions. The simulation outcome indicated that the HFE-ANFIS model has offered superior performance to other methods with the maximum average sensitivity of 94.55%, specificity of 96.41%, precision of 94.66% and accuracy of 95.97%.

Neural network’s selection of color in UI design of social software

In recent years, the design of social software UI has become a design research focus in the field of design. Color affects many factors in UI design. However, there is currently no suitable method for effectively selecting colors in social software. In this paper, the color of social software UI design based on BP neural network is selected. The traditional BP neural network (BP), genetic algorithm improved BP neural network (GA-BP) and Mind Evolution Algorithm improved BP neural network (MEA-BP) are analyzed and summarized. Finally, the strong predictor and thought evolution method are used to improve MEA-BP-Adaboost. The experiment proves that the training results of the MEA-BP-Adaboost neural network are very good, and the color difference is reduced by about 30, 26.5 and 35.3%, respectively, compared to the three different BP neural networks. The color selection method based on MEA-BP-Adaboost can more effectively improve the accuracy of color selection in the UI design of social software, while reducing the number of experiments. In the color selection algorithm, the color accuracy rate and recall rate of the seven different colors are basically between 90 and 95%, which can basically achieve the desired effect. This also proves that the usability of BP neural network in social software UI design is very high. The methods involved in this article can be applied to other color space conversion and other image acquisition, display, processing and output devices. It is believed that these research works have certain theoretical guiding significance and practical application value to promote the development of color image color restoration technology.

Improving Color Space Conversion for Camera-Captured Images via Wide-Gamut Metadata

Color space conversion is the process of converting color values in an image from one color space to another. Color space conversion is challenging because different color spaces have different sized gamuts. For example, when converting an image encoded in a medium-sized color gamut (e.g., AdobeRGB or Display-P3) to a small color gamut (e.g., sRGB), color values may need to be compressed in a many-to-one manner (i.e., multiple colors in the source gamut will map to a single color in the target gamut). If we try to convert this sRGB-encoded image back to a wider gamut color encoding, it can be challenging to recover the original colors due to the color fidelity loss. We propose a method to address this problem by embedding wide-gamut metadata inside saved images captured by a camera. Our key insight is that in the camera hardware, a captured image is converted to an intermediate wide-gamut color space (i.e., ProPhoto) as part of the processing pipeline. This wide-gamut image representation is then saved to a display color space and saved in an image format such as JPEG or HEIC. Our method proposes to include a small sub-sampling of the color values from the ProPhoto image state in the camera to the final saved JPEG/HEIC image. We demonstrate that having this additional wide-gamut metadata available during color space conversion greatly assists in constructing a color mapping function to convert between color spaces. Our experiments show our metadata-assisted color mapping method provides a notable improvement (up to 60% in terms of E) over conventional color space methods using perceptual rendering intent. In addition, we show how to extend our approach to perform adaptive color space conversion based spatially over the image for additional improvements.

A new perspective on decolorization: feature-preserving decolorization

Information loss is a major problem in decolorization for color images. For example, the color-to-gray conversion may result in degradation in the contrast that affects visual perception quality. Preserving the source of information as much as possible is the main goal of the decolorization. This paper introduces an efficient contrast preservation method for decolorization based on spatial statistical distributions of the pixel-pair values. The statistical distributions are represented by the co-occurrence matrix in a compact form. Then, a feature extraction step is carried out by making use of this matrix. The feature extraction process is carried out for each channel of the color image and grayscale image to obtain source and target features. Feature-preserving criterion is constructed by $$l_2$$ norm-based quality metric between source feature and target feature. The proposed method is remarkable because it is adaptable to any feature preserving, such as contrast. Moreover, there are no optimization phase, color space conversion, high complexity, and local mapping in the proposed method. Experimental results show that the performance of the proposed method is comparable to the existing decolorization approaches.

Prediction of Tetralogy of Fallot using Fuzzy Clustering

Background: Congenital Heart Disease is one of the abnormalities in your heart's structure. To predict the tetralogy of fallot in a heart is a difficult task. Cluster is the collection of data objects, which are similar to one another within the same group and are different from the objects in the other clusters. To detect the edges, the clustering mechanism improve its accuracy by using segmentation, Colour space conversion of an image implemented in Fuzzy c-Means with Edge and Local Information. Objective: To predict the tetralogy of fallot in a heart, the clustering mechanism is used. Fuzzy c-Means with Edge and Local Information gives an accuracy to detect the edges of a fallot to identify the congential heart disease in an efficient way. Methods: One of the finest image clustering methods, called as Fuzzy c-Means with Edge and Local Information which will introduce the weights for a pixel value to increase the edge detection accuracy value. It will identify the pixel value within its local neighbor windows to improve the exactness. For evaluation , the Adjusted rand index metrics used to achieve the accurate measurement. Results: The cluster metrics Adjusted rand index and jaccard index are used to evaluate the Fuzzy c- Means with Edge and Local Information. It gives an accurate results to identify the edges. By evaluating the clustering technique, the Adjusted Rand index, jaccard index gives the accurate values of 0.2, 0.6363, and 0.8333 compared to other clustering methods. Conclusion: Tetralogy of fallot accurately identified and gives the better performance to detect the edges. And also it will be useful to identify more defects in various heart diseases in a accurate manner. Fuzzy c-Means with Edge and Local Information and Gray level Co-occurrence matrix are more promising than other Clustering Techniques.

Truncated Multiplier with Delay-Minimized Exact Radix-8 Booth Recoder Using Carry Resist Adder

The delay owing to the generation of odd multiples $$(\pm \,3)$$ in Radix-8 Booth recoding is minimized in this paper using carry resist adder (CRA). CRA is intentionally developed for performing the exact addition of $$\pm \,1$$ and $$\pm \,2$$ without carry propagation. The theoretical delay analysis proves that the 8-bit CRA reduces 86.26% of delay when compared to the conventional Carry Propagate Addition (CPA) methods. Subsequently, the relative comparisons of CRA with various approximation-based recoding show that the CRA consumes fewer area, power and critical path delay. Further, the $$8\times 8$$ and $$16\times 16$$ signed binary multiplication using CRA-based Radix-8 Booth recoder is developed and synthesized on TSMC 65nm CMOS standard cell library. Also, the trade-off between area, power, delay and accuracy is verified for the proposed design using truncation. Finally, the CRA-based truncated Radix-8 Booth $$8\times 8$$ multiplier is applied to the color space conversion for quantifying its amicability in imaging. The PSNR and MSE are used to evaluate the quality of the resultant image and show better performance than other existing approximated as well as truncated Radix-8 Booth multipliers.

Real-time License plate number detection based on image contour

The license plate number detection technology is an application of computer video image recognition technology in vehicle license plate recognition. In this paper, an image processing technology is used to extract and detect a license plate in motion from a complex background to a license plate recognition system. The method in this paper is mainly divided into four steps. First, the original image is preprocessed, the color image is converted into a grayscale image, and the gradient edge features are extracted by Sobel operator. Secondly, after the image is binarized, the image generated through the morphological operation of the image to highlight the image outline. Then detect the contour of the entire image and label the contour. Finally, the RGB space is converted into the HSV color space through color space conversion, and the color of the license plate is used to find the area where the license plate is located in the image.

An Acquisition Method for Visible and Near Infrared Images from Single CMYG Color Filter Array-Based Sensor.

Near-infrared (NIR) images are very useful in many image processing applications, including banknote recognition, vein detection, and surveillance, to name a few. To acquire the NIR image together with visible range signals, an imaging device should be able to simultaneously capture NIR and visible range images. An implementation of such a system having separate sensors for NIR and visible light has practical shortcomings due to its size and hardware cost. To overcome this, a single sensor-based acquisition method is investigated in this paper. The proposed imaging system is equipped with a conventional color filter array of cyan, magenta, yellow, and green, and achieves signal separation by applying a proposed separation matrix which is derived by mathematical modeling of the signal acquisition structure. The elements of the separation matrix are calculated through color space conversion and experimental data. Subsequently, an additional denoising process is implemented to enhance the quality of the separated images. Experimental results show that the proposed method successfully separates the acquired mixed image of visible and near-infrared signals into individual red, green, and blue (RGB) and NIR images. The separation performance of the proposed method is compared to that of related work in terms of the average peak-signal-to-noise-ratio (PSNR) and color distance. The proposed method attains average PSNR value of 37.04 and 33.29 dB, respectively for the separated RGB and NIR images, which is respectively 6.72 and 2.55 dB higher than the work used for comparison.