Hue , Saturation And Value Research Articles

Sunfa Ata Zuyan machine learning models for moon phase detection: algorithm, prototype and performance comparison

The history recorded moon as the most inspiring object in the sky, but it combined with visibility issues to study the phases. This research paper proposes a novel algorithm named Sunfa Ata Zuyan (SAZ), which is meant to extend the shape detection algorithms to aim for lunar phase deceleration and overcome the difficulties encountered by the previous methods to find the moon and determine its phase. The paper sets to investigate two aims. First, propose the add-on algorithm SAZ to determine the lunar phase's data faster. Secondly, evaluate the Raspberry Pi as the main CPU due to its compact size and power as the primary processor based on the idea of a portable designed system. Then to examine the ability of the SAZ algorithm, it's combined with famous algorithms like hue, saturation and value (HSV), Canny, erosion, shape detection, and binarization has been tested on both personal computers (PC) and Raspberry Pi with the same images being compared. The results show that SAZ will help the shape detection algorithm to find the object and disclose the moon phases. Furthermore, the Raspberry Pi, functioning as a CPU, can perform as a hand-to-hand system to determine the lunar phase as a compact portable remote sensing structure.

A new and accessible instrumentation to determine urea in UHT milk using digital image analysis

This research demonstrates the development, optimization and application of a new low-cost detection system, based on digital image analysis, for the detection of urea in ultra-high-temperature (UHT) milk samples. The apparatus built in the laboratory, allows the capture of images through a simple system built by polyvinyl chloride (PVC) tubes, a digital microscope and a peristaltic mini-pump, after the colorimetric reaction between urea and diacetylmonoxime (butane-2,3-dionammonoxime). The red, green and blue (RGB) and hue, saturation and value (HSV) color systems were studied, with the saturation channel of the HSV color system selected as the analytical signal. Subsequently, the experimental chemical conditions were evaluated through multivariate experimental designs and the optimal conditions were defined. The proposed method was validated, and the detection and quantification limits presented by the method were 0.35 mg L−1 and 0.52 mg L−1, respectively; precision, ranged between 1.6 and 2.8 %. The results were compared with those obtained using the mid-infrared technique and no statistically significant differences were observed at a 95 % confidence level. The proposed method was applied to eight UHT milk samples that presented urea content ranging from 187 to 386 mg L−1. The mean values obtained are in agreement with values presented in other studies for the determination of urea in milk. The results indicated that the system described and validated here is promising and can be applied to assess the authenticity and quality of milk.

Mobile visual search model for Dunhuang murals in the smart library

PurposeThe purpose of this paper is to build a mobile visual search service system for the protection of Dunhuang cultural heritage in the smart library. A novel mobile visual search model for Dunhuang murals is proposed to help users acquire rich knowledge and services conveniently.Design/methodology/approachFirst, local and global features of images are extracted, and the visual dictionary is generated by the k-means clustering. Second, the mobile visual search model based on the bag-of-words (BOW) and multiple semantic associations is constructed. Third, the mobile visual search service system of the smart library is designed in the cloud environment. Furthermore, Dunhuang mural images are collected to verify this model.FindingsThe findings reveal that the BOW_SIFT_HSV_MSA model has better search performance for Dunhuang mural images when the scale-invariant feature transform (SIFT) and the hue, saturation and value (HSV) are used to extract local and global features of the images. Compared with different methods, this model is the most effective way to search images with the semantic association in the topic, time and space dimensions.Research limitations/implicationsDunhuang mural image set is a part of the vast resources stored in the smart library, and the fine-grained semantic labels could be applied to meet diverse search needs.Originality/valueThe mobile visual search service system is constructed to provide users with Dunhuang cultural services in the smart library. A novel mobile visual search model based on BOW and multiple semantic associations is proposed. This study can also provide references for the protection and utilization of other cultural heritages.

Stereo-vision-based multi-crop harvesting edge detection for precise automatic steering of combine harvester

With the development of the intelligent combine harvesters, the whole-field automatic harvesting has become feasible and includes driving straight along the unharvested crop edge and steering in the crop headland. Detecting the unharvested crop edge and the end of the cultivation crop, is the foundation of whole-field automatic harvesting. A rapid and robust detection method for detecting the harvesting edge for multiple crops based on stereo vision was developed. In order to improve the processing speed and keep the crop harvesting edge within the target region, a dynamic ROI (region of interest) extraction algorithm based on HSV (hue, saturation, and value) space scanning was developed. A coordinate transformation method was applied to generate a crop elevation image, and the unharvested crop area was obtained by the Ostu algorithm. Simultaneous detection of unharvested crop edge and crop end edge was realised by the method based on the horizontal length feature of the unharvested area contour. This enabled automatic driving along the unharvested crop edge, and automatic steering into the next harvest path according to the crop end edge. Finally, field tests with crops of paddy, oil-seed rape, and maize crops were carried out to test the system performance, and the accuracy of the detection was found to be higher than 98% for paddy and rape, and higher than 94% for maize with an average processing speed of 49 ms frame−1. The experiments showed that the proposed method exhibited high accuracy and efficiency and was suitable for various complex field environments.

An Intelligent Classification System for Trophozoite Stages in Malaria Species

Malaria is categorised as a dangerous disease that can cause fatal in many countries. Therefore, early detection of malaria is essential to get rapid treatment. The malaria detection process is usually carried out with a 100x magnification of thin blood smear using microscope observation. However, the microbiologist required a long time to identify malaria types before applying any proper treatment to the patient. It also has difficulty to differentiate the species in trophozoite stages because of similar characteristics between species. To overcome these problems, a computer-aided diagnosis system is proposed to classify trophozoite stages of Plasmodium Knowlesi (PK), Plasmodium Falciparum (PF) and Plasmodium Vivax (PV) as early species identification. The process begins with image acquisition, image processing and classification. The image processing involved contrast enhancement using histogram equalisation (HE), segmentation procedure using a combination of hue, saturation and value (HSV) color model, Otsu method and range of each red, green and blue (RGB) color selections, and feature extraction. The features consist of the size of infected red blood cell (RBC), brown pigment in the parasite, and texture using Gray Level Co-occurrence Matrix (GLCM) parts. Finally, the classification method using Multilayer Perceptron (MLP) trained by Bayesian Rules (BR) show the highest accuracy of 98.95%, rather than Levenberg Marquardt (LM) and Conjugate Gradient Backpropagation (CGP) training algorithms.

Image Masking and Enhancement System for Melanoma Early Stage Detection

Early stage melanoma detection (ESMD) is crucial as late detection kills. Computer aided diagnosis systems (CADS) integrated with high level algorithms are major tools capable of ESMD with high degree of accuracy, specificity, and sensitivity. CADS use the image and the information within the pixels of the image. Pixels’ characteristics and orientations determine the colour and shapes of the images as the pixels and associated environment are closely interrelated with the lesion. CADS integrated with Convolutional Neural Networks (CNN) specifically play a major role for ESMD with high degree of accuracy. The proposed system has two steps to produce high degree of accuracy; the first is the Fuzzy-Logic (FL) based maskProd and pixel Enhancement vectors and the second is integration with deep CNN algorithms. The vectors, maskProd and pixel Enhancement, are based on lesion masking and image enhancement via fuzzy logic as hSvMask production has three inputs; sChannel, entropyFilter and distanceVector. The originality of this paper is based on the regional enhancement of each pixel of the image by fuzzy-logic. To change the brightness of an image by the conventional methods (CM), it is required to multiply the value of all and each pixels of the image by the same constant or weight value in order to generate new values of the pixels. But in the proposed method, FL determines a special and specific constant or weight value for every single pixel of the image; called a variable weight value. Secondly via conventional or classical methods of image processing, mask production is realized based on the threshold value predetermined, and this threshold value decides whether the image lesion is melanoma or not. CM use constant weight values but in the method suggested, FL determines a specific variable weight value for every single pixel of the image. Secondly CMs realize the mask production using predetermined threshold values, but proposed method suggests a new way capable of detecting pixels in the lesion-free regions, to do that, a system using FL was developed for each and every single pixel via the data obtained from three sources: the s-channel of the HSV (Hue, Saturation and Value) colour space, distance vector, and image entropy vector. The proposed method and its integration with algorithms; such as AlexNet, GoogLeNet, Resnet18, VGG-16, VGG-19, Inception-V3, ShuffleNet and Xception, detected the melanoma with an accuracy of 86.95, 85.60, 86.45, 86.70, 85.47, 86.76, 86.45 and 85.30 percentages, respectively.

Evaluation of tree-based ensemble algorithms for predicting the big five personality traits based on social media photos: Evidence from an Iranian sample

Machine learning has been widely used in psychological issues. In this study, Iranian Instagram users' photos were used to predict the big five personality traits with extra trees, gradient boosted trees and random forest. The big five personality traits of 142 users were measured using the short form of Goldberg's 50-item personality scale, and 30 color features including classified color histogram, color level mean, and color level median in RGB (Red, Green, and Blue) and HSV (Hue, Saturation, and Value) spaces were extracted from the 1348 photos. Predictive models were cross-validated, and the highest root mean square error and mean absolute error values for all models did not reach 0.16 and 0.13, respectively, which were related to the prediction of Extraversion. Overall, all models had good performance and generalizability in prediction. Results showed that each personality trait has a special type of color preference and photo posting. The features' importance was evaluated by the Gini method. Color histogram and HSV color space were the most important features in modeling. As a result, it can be said that the proposed methodology is effective in practice, but care must be taken with multicollinearity, accuracy, and integrity of social media data.

Kernelised correlation filters target tracking fused multi‐feature based on the unmanned aerial vehicle platform

As unmanned aerial vehicle (UAV) emerged as a flexible acquisition system that is widely used in military and civilian fields, efficient target tracking algorithm is in urgent need for UAV-based computer vision. Although research studies have been reported on typical interferences in the tracking process such as scale change, occlusion, distortion etc., some issues still exist for the target tracking algorithm based on UAV vision. This study exploited the features hidden in different colour spaces, and proposed a multi-feature multi-filter fusion tracking method that combines the HSV (hue, saturation and value) colour space with the histogram of oriented gradient (HOG) feature. The HSV colour space is proved to be able to discriminate objects under different conditions. The HOG of each HSV channel is utilised to train Kernelised correlation filters (KCF), respectively. The final tracking result is the candidate result with the biggest peak sidelobe ratio (PSR). Computer simulations proved that the fusion strategy proposed in this study can effectively improve the tracking performance of the tracker especially when the image sequences are interfered by deformation, occlusion, low resolution, etc. The performance of the tracker is also tested on UAV.

Luminosity and contrast enhancement of retinal vessel images using weighted average histogram

Retinal imaging can detect a variety of common eye illnesses as well as cardiovascular diseases. At the same time, poor-quality retinal images are impossible to diagnose due to non-uniform illumination, image obscuration, and low contrast. A novel enhancement method is proposed to improve the luminosity and contrast of the color retinal image. This algorithm is applied to the publicly available structured analysis of the retina (STARE) image dataset, and promising results are obtained. Contrast enhancement and luminosity enhancement are the two sections of the proposed method. Gamma adjustment on the V channel in hue, saturation and value (HSV) color model gives a luminosity gain matrix, which improves the image luminosity. Weighted average histogram equalization (WAHE) acts on the luminosity channel (L) of the L*a*b* (luminosity and chromaticity) model for contrast improvement. Finally, an effective luminosity and contrast adjustment scheme is presented for various retinal vessel images, and its performance is illustrated against the different state-of-the-art methods using a well-known numerical metric.

Prototyping Digital Tongue Diagnosis System on Raspberry Pi

Tongue inspection is a complementary diagnosis method that widely used in Traditional Chinese Medicine (TCM) by inspecting the tongue body constitution to decide the physiological and pathological functions of the human body. Since tongue manifestation is done by practitioner’s observation using naked eye, many limitations can affect the diagnosis result including environment condition and experiences of the practitioner. Lately, tongue diagnosis has been widely studied in order to solve these limitations via digital system. However, most of recent digital system are bulky and not equipped with intelligent diagnosis system that can finally predict the health status of the patient. In this research, digital tongue diagnosis system that uses intelligent diagnosis consisted of image segmentation analysis, tongue coating recognition analysis, and tongue color classification has been embedded on Raspberry Pi. Tongue segmentation implements Hue, Saturation and Value (HSV) color space with Brightness Conformable Multiplier (BCM) for adaptive brightness filtering to recognized tongue body accurately while eliminating perioral area. Tongue Coating Recognition uses threshold method to detect tongue coating and eliminate the unwanted features including shadow. Tongue color classification uses hybrid method consisted of k-means clustering and Support Vector Machine (SVM) to classify between red, light red and deep red tongue and further gave diagnosis based on color. This experiment concluded that it is feasible to embed the algorithm on Raspberry Pi to promote system portability while attaining similar accuracy for future telemedicine.

Gabor Features Extraction and Land-Cover Classification of Urban Hyperspectral Images for Remote Sensing Applications

The principles of the transform stage of the extract, transform and load (ETL) process can be applied to index the data in functional structures for the decision-making inherent in an urban remote sensing application. This work proposes a method that can be utilised as an organisation stage by reducing the data dimension with Gabor texture features extracted from grey-scale representations of the Hue, Saturation and Value (HSV) colour space and the Normalised Difference Vegetation Index (NDVI). Additionally, the texture features are reduced using the Linear Discriminant Analysis (LDA) method. Afterwards, an Artificial Neural Network (ANN) is employed to classify the data and build a tick data matrix indexed by the belonging class of the observations, which could be retrieved for further analysis according to the class selected to explore. The proposed method is compared in terms of classification rates, reduction efficiency and training time against the utilisation of other grey-scale representations and classifiers. This method compresses up to 87% of the original features and achieves similar classification results to non-reduced features but at a higher training time.

A machine learning model pipeline for detecting wet pavement condition from live scenes of traffic cameras

Highway safety is largely influenced by weather conditions that have become increasingly volatile due to the climate change. It well known that wet pavement significantly reduces surface friction, leading to inflated collision risk. Thus, timely knowledge of the road surface condition is critical for safe driving. In this paper, a novel machine learning model pipeline is proposed to detect the wetness of pavement based on live images of highway scenes captured by publicly accessible traffic cameras. To simplify the learning task, we finetuned the state-of-the-art instance segmentation baseline models to extract background instance targets, including pavement, sky, and vegetation, which are common in highway scenes. Then, the color mixture attributes in HSV (hue, saturation and value) of each segmented instance were extracted and used as visual cues for inferring pavement condition. Finally, gradient boosting ensemble classifiers are constructed and trained using the HSV features to predict the wetness of pavement. For the segmentation task, we leveraged Detectron2 baseline models (Mask R-CNN) and evaluated three backbone networks: R50-FPN, R101-FPN, and X101-FPN. For the classification task, two most popular gradient boosting algorithms (XGBoost and CatBoost) were evaluated together with a classic logistic model. Based on experiments with our custom dataset, the best performance (F1 score: 0.927, AUC: 0.975) was achieved by the R101-FPN backbone coupled with the CatBoost classifier.

Online Monitoring Strategies for Colorimetric Detection of Cadmium Ions and pH Based on Gold Nanomaterials with a Low-Cost Color Sensor

Gold triangular nanoplates (AuTNPs) and gold nanorods (AuNRs) have been widely used in colorimetric sensing due to their shape-dependent localized surface plasmon resonance properties. Herein, we report new colorimetric detection strategies based on a low-cost RGB colorimetry system equipped with an industrial-grade color sensor and provide two demonstrations for the potential applications. The first is the online detection of cadmium ions (Cd²⁺) through the Cd²⁺-promoted etching of glutathione-stabilized AuTNPs that was initialized by H₂O₂ and iodide. For the first time, the etching time corresponding to the inflection point of the first derivative of the Red channel transmission curve was used for calibration. The method was finally applied in the detection of Cd²⁺ in fortified rice samples. In the second demonstration, the RGB values continuously acquired using the instrument were converted to HSV (hue, saturation, and value) space, which was successfully employed in pH detection via etching of AuNRs. Importantly, the overaged AuNRs with poor homogeneity could be discerned by the method at the beginning of analysis, which is helpful in achieving reliable measurements. Our results suggest that integration of the low-cost portable instrument and the new strategies has the promising potential for sensitive, accurate, precise, and good-linearity detection.

Predicting sensory evaluation of spinach freshness using machine learning model and digital images.

The visual perception of freshness is an important factor considered by consumers in the purchase of fruits and vegetables. However, panel testing when evaluating food products is time consuming and expensive. Herein, the ability of an image processing-based, nondestructive technique to classify spinach freshness was evaluated. Images of spinach leaves were taken using a smartphone camera after different storage periods. Twelve sensory panels ranked spinach freshness into one of four levels using these images. The rounded value of the average from all twelve panel evaluations was set as the true label. The spinach image was removed from the background, and then converted into a gray scale and CIE-Lab color space (L*a*b*) and Hue, Saturation and Value (HSV). The mean value, minimum value, and standard deviation of each component of color in spinach leaf were extracted as color features. Local features were extracted using the bag-of-words of key points from Oriented FAST (Features from Accelerated Segment Test) and Rotated BRIEF (Binary Robust Independent Elementary Features). The feature combinations selected from the spinach images were used to train machine learning models to recognize freshness levels. Correlation analysis between the extracted features and the sensory evaluation score showed a positive correlation (0.5 < r < 0.6) for four color features, and a negative correlation (‒0.6 < r < ‒0.5) for six clusters in the local features. The support vector machine classifier and artificial neural network algorithm successfully classified spinach samples with overall accuracy 70% in four-class, 77% in three-class and 84% in two-class, which was similar to that of the individual panel evaluations. Our findings indicate that a model using support vector machine classifiers and artificial neural networks has the potential to replace freshness evaluations currently performed by non-trained panels.

Tracking of moving athlete from video sequences using flower pollination algorithm

Performance analysis, as related to sport, is a process underpinned by a systematic analysis of information, to accelerate the performance of athletes through crafted focused practice session based on the obtained analysis. Quantification of athlete performance profile using sports video has thus been put forward, where the athlete tracking in such video-based analysis is one of the critical elements for the success of an object tracking system. In this study, for the first time the flower pollination algorithm (FPA) is utilised to track the motion of the moving athlete from the sports video. Initially, a search window with the attributes of centroid coordinates of the moving athlete, width and length of the search window is used to represent the current position of the athlete. Subsequently, the hue, saturation and value (HSV) histogram of the region within the search window is evaluated. In the consecutive frame, several potential positions of the athlete are identified, and the Bhattacharyya distance between the HSV histogram of the athlete in the previous frame and the potential position in the current frame is calculated. Since the FPA attempts to maximise the similarity of both histograms, intuitively, the current position of the moving athlete should be only slightly different than his previous position. The comparative analysis shows that the FPA is comparable with other competing algorithms in terms of detection rate, tracking accuracy and processing time.

Recognition of feeding behaviour of pigs and determination of feeding time of each pig by a video-based deep learning method

Monitoring the feeding behaviour of pigs and measuring their feeding time can help farmers to evaluate the pig health and welfare. The aim of this study was to develop a video-based deep learning algorithm to recognise feeding behaviour of nursery pigs and determine the feeding time of pigs on an individual level. In the experiment, two pens of pigs were video recorded for 3 days. In the video of feeding region of pen 1, 96,000 1 s feeding episodes and 96,000 1 s non-feeding episodes were generated. 70% of these data was randomly selected as training set and the remaining 30% as validation set. In the video of feeding region of pen 2, all the labelled 24,656 1 s feeding episodes and 61,744 1 s non-feeding episodes were used as test set. Firstly, the Convolutional Neural Network (CNN) architecture Xception was used to extract spatial features. These features were input into Long Short-term Memory (LSTM) framework to extract spatial-temporal features. Through the fully connected layer, the prediction function Softmax was finally used to classify these 1 s episodes as feeding or non-feeding. An image processing algorithm based on maximum entropy segmentation, HSV (Hue, Saturation and Value) colour space transformation and template matching was proposed to calculate the circularity of the head, the ratio of the head to the feeding sub-region, the accumulated pixels of the head motion, and the distance from the head to the number on pig back in order to determine the identity and feeding time of each pig. In the test set, the proposed algorithm could recognise feeding behaviour with an accuracy of 98.4%, a sensitivity of 98.8%, specificity of 98.3% and precision of 95.9%, and could correctly recognise 98.5% of feeding time of individual pigs from the total 45944 s feeding time of 8 pigs. The results indicate that the proposed method can be used to recognise feeding behavior of pigs and determine feeding time of each pig.

Application of Color Featuring and Deep Learning in Maize Plant Detection

Maize plant detection was conducted in this study with the goals of target fertilization and reduction of fertilization waste in weed spots and gaps between maize plants. The methods used included two types of color featuring and deep learning (DL). The four color indices used were excess green (ExG), excess red (ExR), ExG minus ExR, and the hue value from the HSV (hue, saturation, and value) color space, while the DL methods used were YOLOv3 and YOLOv3_tiny. For practical application, this study focused on performance comparison in detection accuracy, robustness to complex field conditions, and detection speed. Detection accuracy was evaluated by the resulting images, which were divided into three categories: true positive, false positive, and false negative. The robustness evaluation was performed by comparing the average intersection over union of each detection method across different sub–datasets—namely original subset, blur processing subset, increased brightness subset, and reduced brightness subset. The detection speed was evaluated by the indicator of frames per second. Results demonstrated that the DL methods outperformed the color index–based methods in detection accuracy and robustness to complex conditions, while they were inferior to color feature–based methods in detection speed. This research shows the application potential of deep learning technology in maize plant detection. Future efforts are needed to improve the detection speed for practical applications.

Welding flame detection based on color recognition and progressive probabilistic Hough transform

SummaryWelders are widely used in machinery manufacturing and machining industry. In the process of electric welding, accidents are often caused by some unsafe behaviors. This paper focuses on the intelligent recognition method of welding flame, which can reduce the accidents of welding flame to a certain extent. Due to the difficulties in observing workers on job sites, developing an algorithm for intelligent identification of unsafe factors is significant. Firstly, Hue,Saturation and Value (HSV) color space with strong robustness is selected to segment the mark image with threshold value, and the region of interest in which the welding fame is located is segment; secondly, the Progressive Probabilistic Hough transform is used to detect the region of interest, judge the number of existing lines, and finally determine whether there is welding flame in this region. In this way, the welding flame can be better detected, which can reduce the workload of managers and is beneficial to the supervision of the management personnel.

Cucumber disease detection using adaptively regularised kernel-based fuzzy C-means and probabilistic neural network

India is an agricultural country. The major position of India depends on agriculture. But due to diseases in leaves, there is a great loss to farmers. To avoid this problem, automatic disease detection of cucumber disease is proposed. The proposed methodology consists of three modules namely, segmentation, feature extraction cucumber disease detection. Initially, the cucumber diseases are segmented using adaptively regularised kernel-based fuzzy C-means (ARKFCM). Once the disease is segmented, the colour features are extracted using hue, saturation and value (HSV) technique and texture features are extracted using grey level co-occurrence matrix (GLCM) technique. After the feature extraction process, the extracted features are given to probabilistic neural network (PNN) to recognise the image as anthracnose, downy mildew and grey mould. Finally, the experimental results demonstrate that our method is efficient and powerful to recognise the cucumber diseased image and its performance is analysed in terms of accuracy, sensitivity and specificity.

Improved GMM‐based method for target detection

The Gaussian mixture model (GMM) is prone to large-scale misdetection in the static case where the background and foreground have similar colours. This study presents an improved GMM method to solve this problem. First, the principal component analysis is used to transform the high-dimensional space into the low-dimensional one with three colour channels, which aims to reduce runtime. Then, the images are processed by GMM to obtain the foreground areas. At the same time, the mean and difference of pixel features in red, green and blue and hue, saturation and value (HSV) colour models are calculated to analyse the similar-coloured components. The areas with similar colours are classified in this process, and thus the foreground is separated from similar-coloured background. Then, all of the extracted areas are targets. The experimental results show that the presented method identified the target areas with high accuracy and high efficiency. Compared with conventional GMM, the presented method obtained the detection-accuracy improvement by about 6% on average.