Machine vision approach for classification of citrus leaves using fused features

In this study emotion-based face expression recognition framework has been proposed using a machine vision (MV) approach. The face emotion dataset has been collected local survey in Bahawalpur city dataset divide into 3 classes happy, sad, and angry. A total of 600 images of size (256 x 256) were transformed into a gray level format and employed a median filter for noise removal. Three non-overlapping regions of interest (ROIs) of size (50 x 50) have been taken and analyze 1800 (600 x 3) ROIs on the overall dataset. Total 45 Statistical features named as texture, histogram, and binary features were extracted. Select optimize features using the correlation-based feature section technique. The optimized dataset employed of MV classifiers namely random forest (RF), logistic (Lg), and J48 are obtained very promising accuracy 96.33%, 95.67%, and 95.33% respectively.

In this study, we describe a technique that used a machine learning (ML) approach to classify four (4) different cotton leaf varieties namely; BS-15, S-32, Z-31, and Z-32. Each variety of cotton leaves were collected from 500 Farmers. These image datasets are captured by using the cell phone camera in the open agricultural field area, and every image was captured from both sides (Front and Back) of the cotton leaf. Each variety of cotton has used over 300 (150 Front Side and 150 Back Side of the leaves) leaf images and the total calculated cotton leaves are 1200 (300 x 4) as leaf image samples. These sample datasets have analyzed through image preprocessing and image segmentation process. Each image was employing four different non-over-lapping regions of interest (ROI’s) and calculated a total of 4800 (1200 x 4) ROI’s. The acquired datasets are employed different machine learning features such as Scalability, Texture, Spectral, Binary, Histogram, Rotational, and translational (R-S-T). A total of fifty-seven (57) machine learning features were evaluated on each ROI and a total calculated 273,600 (4800 x 57) features. Furthermore, the Correlation-Based Feature Selection (CFS) genetic algorithm technique was employed for feature optimization. It has been evaluated 22 optimized features and applying different machine learning (M-L) classifiers namely; K-Nearest Neighbor (K-NN), K*, Random Forest (RF) Tree, and Naive Bayes (NB) Tree. The resulting accuracy produced by K-NN presented is 98.9167% on (512 x 512) ROI’s. The individually overall result accuracy dataset values by using K-NN classifier on the four varieties of cotton leaf namely; BS-15, S-32, Z-31, and Z-32 were evaluated 97.83%, 99.50%, 99%, and 99.33%, respectively.

It is highly acknowledged in the medical profession that density of breast tissue is a major cause for the growth of breast cancer. Increased breast density was found to be linked with an increased risk of breast cancer growth, as high density makes it difficult for radiologists to see an abnormality which leads to false negative results. Therefore, there is need for the development of highly efficient techniques for breast tissue classification based on density. This paper presents a hybrid scheme for classification of fatty and dense mammograms using correlation-based feature selection (CFS) and sequential minimal optimization (SMO). In this work, texture analysis is done on a region of interest selected from the mammogram. Various texture models have been used to quantify the texture of parenchymal patterns of breast. To reduce the dimensionality and to identify the features which differentiate between breast tissue densities, CFS is used. Finally, classification is performed using SMO. The performance is evaluated using 322 images of mini-MIAS database. Highest accuracy of 96.46% is obtained for two-class problem (fatty and dense) using proposed approach. Performance of selected features by CFS is also evaluated by Naïve Bayes, Multilayer Perceptron, RBF Network, J48 and kNN classifier. The proposed CFS-SMO method outperforms all other classifiers giving a sensitivity of 100%. This makes it suitable to be taken as a second opinion in classifying breast tissue density.

This research aims at optimizing the monitoring network used to consistently identify pollution’s origin in the pollution source identification problem in groundwater hydraulics under real-time/operational applications. For this task, Machine Learning (ML) and Deep Learning (DL) methods are introduced, which can outperform metaheuristics, such as Genetic Algorithms (GAs), in terms of total computational load. To test the approach, a theoretical aquifer with two pumping wells is studied, where one of six possible pollution sources may spread a conservative pollutant. An existing own software simulates a 2D surrogate steady state flow field, using particle tracking to simulate advective mass transport only. A large number of combinations of possible source locations (4 different layout scenarios), hydraulic gradients and pumping wells’ flow-rates is used to calculate various features (such as pollutant arrival times, hydraulic drawdowns) in a 29 × 29 grid. Three ML/DL methods (Random Forests, Multi-Layer Perceptron, Convolutional Neural Networks) are tested for prediction accuracy, while Correlation based Feature Selection (CFS), and targeted tests are used to select subsets/sampling frequencies that can provide similar accuracy with the full datasets. This evaluation process bears promising results and paves the way for monitoring network optimization.KeywordsMachine LearningSource IdentificationMonitoring network

The accurate preoperative histopathological grade diagnosis of adult gliomas is of great significance for the formulation of a surgical plan and the implementation of a subsequent treatment. The aim of this study is to establish a predictive model for classifying adult gliomas into grades 2-4 based on preoperative conventional multimodal MRI radiomics. Patients with pathologically confirmed gliomas at Huashan Hospital, Fudan University, between February 2017 and July 2019 were retrospectively analyzed. Two regions of interest (ROIs), called the maximum anomaly region (ROI1) and the tumor region (ROI2), were delineated on the patients' preoperative MRIs utilizing the tool ITK-SNAP, and Pyradiomics 3.0 was applied to execute feature extraction. Feature selection was performed utilizing a least absolute shrinkage and selection operator (LASSO) filter. Six classifiers, including Gaussian naive Bayes (GNB), random forest (RF), K-nearest neighbor (KNN), support vector machine (SVM) with a linear kernel, adaptive boosting (AB), and multilayer perceptron (MLP) were used to establish predictive models, and the predictive performance of the six classifiers was evaluated through five-fold cross-validation. The performance of the predictive models was evaluated using the AUC and other metrics. After that, the model with the best predictive performance was tested using the external data from The Cancer Imaging Archive (TCIA). According to the inclusion and exclusion criteria, 240 patients with gliomas were identified for inclusion in the study, including 106 grade 2, 68 grade 3, and 66 grade 4 gliomas. A total of 150 features was selected, and the MLP classifier had the best predictive performance among the six classifiers based on T2-FLAIR (mean AUC of 0.80 ± 0.07). The SVM classifier had the best predictive performance among the six classifiers based on DWI (mean AUC of 0.84 ± 0.05); the SVM classifier had the best predictive performance among the six classifiers based on CE-T1WI (mean AUC of 0.85 ± 0.06). Among the six classifiers, based on ROI1, the MLP classifier had the best prediction performance (mean AUC of 0.78 ± 0.07); among the six classifiers, based on ROI2, the SVM classifier had the best prediction performance (mean AUC of 0.82 ± 0.07). Among the six classifiers, based on the multimodal MRI of all the ROIs, the SVM classifier had the best prediction performance (average AUC of 0.85 ± 0.04). The SVM classifier, based on the multimodal MRI of all the ROIs, achieved an AUC of 0.81 using the external data from TCIA. The prediction model, based on preoperative conventional multimodal MRI radiomics, established in this study can conveniently, accurately, and noninvasively classify adult gliomas into grades 2-4, providing certain assistance for the precise diagnosis and treatment of patients and optimizing their clinical management.

Hyperspectral remote sensing data contains a large number of spectral bands along with a large a number of features. So for classification purposes, it becomes imperative to reduce the number of spectral bands as well as to reduce the number of features to achieve higher performance in terms of accuracy and computational complexity. Principal component analysis (PCA) and minimum noise fraction (MNF) were used for dimension reduction of spectral bands, on the other hand correlation based feature selection (CFS) technique was employed for selection of features. Parametric classifier like maximum likelihood classifier (MLC), advanced non parametric classifiers like support vector machine (SVM), Multilayer Perceptron (MLP) and ensemble random forest (RF) classifier were used for investigation in two test datasets of different land cover characteristics of Hyperion sensor. It was observed that all of the classifiers have behaved differently with the dimension reduced dataset produced either in spectral bands or feature level. Most of the classifiers have achieved higher performance with dimension reduced dataset produced by MNF as compared to PCA. On the other hand, dataset with selected features could not give noticeably better classification performance than the datasets with entire features. However, computational complexity of the classifiers has been reduced greatly while they were classified with selected features.

The liver tumor is one of the most foremost critical causes of death in the world. Nowadays, Medical Imaging (MI) is one of the prominent Computer Vision fields (CV), which helps physicians and radiologists to detect and diagnose liver tumors at an early stage. Radiologists and physicians use manual or semi-automated systems to read hundreds of images, such as Computed Tomography (CT) for the diagnosis. Therefore, there is a need for a fully-automated method to diagnose and detect the tumor early using the most popular and widely used imaging modality, CT images. The proposed work focuses on the Machine Learning (ML) methods: Random Forest (RF), J48, Logistic Model Tree (LMT), and Random Tree (RT) with multiple automated Region of Interest (ROI) for multiclass liver tumor classification. The dataset comprises four tumor classes: hemangioma, cyst, hepatocellular carcinoma, and metastasis. Converted the images into gray-scale, and the contrast of images was improved by applying histogram equalization. The noise was reduced using the Gabor filter, and image quality was improved by applying an image sharpening algorithm. Furthermore, 55 features were acquired for each ROI of different pixel dimensions using texture, binary, histogram and rotational, scalability, and translational (RST) techniques. The correlation-based feature selection (CFS) technique was deployed to obtain 20 optimized features from these 55 features for classification. The results showed that RF and RT performed better than J48 and LMT, with an accuracy of 97.48% and 97.08%, respectively. The proposed novel framework will help radiologists and physicians better diagnose liver tumors.

The proliferation of Distributed Denial-of-Service (DDoS) attacks poses critical threats to network infrastructure, while conventional intrusion detection systems struggle to adapt to evolving attack patterns. Although ensemble learning methods achieve high accuracy on benchmark datasets, their opaque decision-making processes hinder deployment in Security Operations Centers (SOCs). To address this interpretability-performance gap, we propose an explainable machine learning framework integrating comparative benchmarking with quantitative interpretability analysis using the CIC-DDoS2019 dataset. Six supervised algorithms Decision Tree, Random Forest, XGBoost, LightGBM, Multilayer Perceptron, and Naïve Bayes were evaluated under standardized preprocessing protocols including random undersampling (50:50 class ratio), correlation-based feature selection (r > 0.9 threshold), and three-tier validation combining hold-out testing, train-validation splits, and 5-fold stratified cross-validation. LightGBM achieved optimal performance with 99.96% accuracy and F1-score of 0.9996, outperforming simple baselines by 0.35 percentage points while demonstrating superior computational efficiency. Beyond conventional performance metrics, we introduce the Feature Stability Score (FSS), a novel quantitative measure of SHAP-based feature importance consistency across validation folds. Spearman correlation analysis reveals a strong positive relationship between FSS and model robustness measured by cross-validation variance (ρ = 0.857, p = 0.014), establishing that stable feature attributions predict superior generalization. SHAP analysis identifies Flow Duration, Bwd Packet Length Mean, Fwd Packet Length Max, and Flow IAT Mean as dominant attack indicators. This integrated framework demonstrates that combining explainable AI with ensemble learning enables accurate, robust, and interpretable DDoS detection suitable for operational cybersecurity deployments.

Osteoarthritis (OA) is a disease that commonly affects the joints in humans. It is the major diseases in aging and obesity society. OA has affected millions of people around the world. OA has affected 10% of Thai population in 2014. According to National Statistical Office (NSO) estimates, 31.8% of Thai population will be affected in OA by 2050. It is difficult to bring back to normal when people suffer from OA. X-ray imaging is the basic method to detect OA. There are 5 grades to differentiate the level of OA which starts from grade 0 to grade 4. This proposed work is to classify the stage of OA by applying image analysis, and classification techniques. The first step of the proposed technique is to identify the region of interest (ROI). Which regions can specify of each texture image. There are four ROIs: (i) Lateral Femur (LF), (ii) Medial Femur (MF), (iii) Lateral Tibia (LT), (iv) Medial Tibia (MT). When ROI was identified, next each ROI (sub-image) was applied with texture descriptor for extractin the feature vectors. To reduce the dimension of feature space, Correlation-based Feature Selection (CFS) approach was implemented. Finally, these selected feature vectors were used to generate the desired classifiers. The research challenge is to find which ROI can produce the better classification result. With respect to a collection of 130 images data set, the obtained result found that Medial Tibia (MT) produced the best performance with Area Under the ROC Curve (AUC) value of 0.871.

Emotion recognition plays a pivotal role in advancing Human-Computer Interaction (HCI) by enabling systems to understand and respond to human emotional states. Among various modalities, Speech Emotion Recognition (SER) stands out as a non-invasive, cost-effective, and temporally efficient approach for detecting emotions. This study leverages the RAVDESS dataset, which provides high-quality audio recordings for emotion classification. The proposed methodology involves preprocessing audio signals to remove noise, extracting temporal and spectral features in both time and frequency domains, and implementing machine learning models for multi-class emotion classification. The study evaluates and compares the performance of several classification models, including Random Forest (RF), Multilayer Perceptron (MLP), Support Vector Machine (SVM), Convolutional Neural Networks (CNN), and Decision Trees (DT). Experimental results demonstrate promising accuracy levels, highlighting the potential of these machine learning techniques in SER. This research contributes to the integration of SER within Brain-Computer Interfaces (BCI) and other emotion-aware applications, paving the way for enhanced interactive systems. Keywords – Speech Emotion Recognition -Machine Learning - Random Forest -Multilayer Perceptron - Support Vector Machine - Convolutional Neural Networks - Decision Tree - RAVDESS Dataset

Abstract. Cloud fraction as a vital component of meteorological satellite products plays an essential role in environmental monitoring, disaster detection, climate analysis and other research areas. Random forest (RF) and multilayer perceptron (MLP) algorithms were used in this paper to retrieve the cloud fraction of AGRI (Advanced Geosynchronous Radiation Imager) on board the Fengyun-4A (FY-4A) satellite based on its full-disk level-1 radiance observation. Corrections have been made subsequently to the retrieved cloud fraction in areas where solar glint occurs using a correction curve fitted with sunglint angle as weight. The algorithm includes two steps: the cloud detection is conducted firstly for each AGRI field of view to identify whether it is clear sky, partly cloudy or overcast within the observation field. Then, the cloud fraction is retrieved for the scene identified as partly cloudy. The 2B-CLDCLASS-lidar cloud fraction product from the CloudSat and CALIPSO active remote sensing satellite is employed as the truth to assess the accuracy of the retrieval algorithm. Comparison with the operational AGRI level-2 cloud fraction product is also conducted at the same time. The results indicate that both the RF and MLP cloud detection models achieved high accuracy, surpassing that of operational products. However, both algorithms demonstrated weaker discrimination capabilities for partly cloudy conditions compared to clear-sky and overcast situations. Specifically, they tended to misclassify fields of view with low cloud fractions (e.g., cloud fraction = 0.16) as clear sky and those with higher cloud fractions (e.g., cloud fraction = 0.83) as overcast. Between the two models, RF exhibited higher overall accuracy. Both RF and MLP models performed well in cloud fraction retrieval, showing lower mean error (ME), mean absolute error (MAE) and root mean square error (RMSE) compared to operational products. The ME for both RF and MLP cloud fraction retrieval models was close to zero, while RF had slightly lower MAE and RMSE than MLP. During daytime, the high reflectance in sunglint areas led to larger retrieval errors for both RF and MLP algorithms. However, after correction, the retrieval accuracy in these regions improved significantly. At night, the absence of visible light observations from the AGRI instrument resulted in lower classification accuracy compared to daytime, leading to higher cloud fraction retrieval errors during nighttime.

PurposeThe purpose of this paper is to describe innovative machine vision methods that have been employed for the capture and analysis of 3D skin textures; and the resulting potential for assisting with identification of suspicious lesions in the detection of skin cancer.Design/methodology/approachA machine vision approach has been employed for analysis of 3D skin textures. This involves an innovative application of photometric stereo for the capture of the textures, and a range of methods for analysing and quantifying them, including statistical methods and neural networks.Findings3D skin texture has been identified as a useful indicator of skin cancer. It can be used to improve realism of virtual skin reconstructions in tele‐dermatology. 3D texture features can also be combined with 2D features to obtain a more robust classifier for improving diagnostic accuracy, thereby assisting with the long‐term goal of implementing computer‐aided diagnostics for skin cancer.Originality/valueThe device developed for capturing 3D skin textures is known as the “Skin Analyser”, and as far as the authors know it is unique in the world in being able to recover 3D textures from pigmented lesions in vivo. There currently exist numerous methods for analysing lesions, including manual inspection (using established heuristics commonly known as ABCD rules), dermoscopy and SIAoscopy. The ability to capture and analyse 3D lesion textures complements these existing techniques and forms a valuable additional indicator for assisting with the early detection of dangerous skin cancers such as melanoma.

Reliable and accurate measurement of particle size and subsequent analysis of particle size distribution (PSD) is central to characterization of particulate minerals. Traditionally PSD of particulate materials was determined using standard mechanical sieving, but it was shown to be deficient in accurate particle size measurements, hence PSD analysis. In this paper a highly inexpensive machine vision based approach was proposed as an alternative to mechanical sieving. The test mineral used was ball-milled celestite. This machine vision approach used a document scanner as imaging device and a user-coded java ImageJ plugin performed the image processing and automated the PSD analysis. The acquired color images were preprocessed to binary images and the particles analyzed after grouping them based on their distinct length. Volumes of all particles were evaluated assuming a prolate spheroid geometry from measured lengths and widths. A new approach of using sum of volumes as weighting factor (Svolume) was utilized for particle grouping in ASABE Standard PSD analysis. The plugin also evaluated several significant dimensional (16) and distributional (27) parameters that characterize the PSD of samples. Standard mechanical sieving of samples was performed for experimental verification and results compared with machine vision method. The cumulative undersize PSD followed a log-normal distribution, and the plot against particle size exhibited a linear trend. Shapes of log-normal plot of cumulative undersize PSD were similar between both methods; however, the mechanical sieving curves lagged by 0.2-0.5 mm. The deviation was attributed to the falling-through effect of longer particles through sieve openings.

Anemia is one of the common adverse reactions after hip fracture surgery. The traditional method to solve anemia is allogeneic transfusion. However, the transfusion may lead to some complications such as septicemia and fever. So far, few studies have reported roles of machine learning in predicting whether blood transfusion is needed or not after hip fracture surgery. Therefore, the purpose of this study is to develop machine learning models to predict the likelihood of postoperative blood transfusion in patients undergoing hip fracture surgery. This study enrolled 1355 patients who underwent hip fracture surgery at the Affiliated Hospital of Qingdao University from January 2016 to December 2021. Among all patients, 210 cases received postoperative blood transfusion. All patients were randomly divided into a training group and a testing group at a ratio of 7:3. In the training group, univariate and multivariate logistic regression analyses were used to determine independent risk factors for the postoperative transfusion. Then, based on these independent risk factors, tenfold cross-validation method was utilized to develop five machine learning models, including logistic, multilayer perceptron (MLP), extreme gradient boosting (XGBoost), random forest (RF), and support vector machine (SVM). The receiver operating characteristic (ROC) curve, area under ROC curve (AUC), and Matthews correlation coefficient (MCC) were generated to evaluate the performance of the models. Calibration plot and decision curve analysis (DCA) were used to test the performance, stability, and clinical applicability of the models. The models were validated using the testing group; and the ROC curve, MCC, calibration plot, and DCA curves were also generated to validate the performance, stability, and clinical applicability of the models. To further verify the robustness of the model, we randomly grabbed 70% of the samples in the testing set, performed 1000 iterations, and calculated the AUC and confidence interval of the five models. Finally, we used SHapley Additive exPlanations (SHAP) to explain these models. Multivariate logistic regression analysis showed that there were 8 independent risk factors, including age, blood transfusion history, albumin (ALB), globulin (GLO), total bilirubin (TBIL), indirect bilirubin (IBIL), hemoglobin (HB), and blood loss > 200ml. We finally selected five independent risk factors including HB, GLO, age, IBIL, and blood loss > 200ml. Based on these five independent risk factors, we generated six characteristic variables, namely HB, HB × HB, HB × blood loss, GLO × HB, age, age × IBIL, and established five machine learning models using a tenfold cross-validation method. In the training group, the AUC values of logistic, RF, MLP, SVM, and XGB were 0.9320, 0.8911, 0.9327, 0.9225, and 0.8825, respectively, and the average AUC was 0.9122 ± 0.0212. The MCC values were 0.65, 0.77, 0.65, 0.66, and 0.68, respectively, and the calibration plot and DCA performed well. In the testing group the AUC values of logistic, RF, MLP, SVM, and XGB were 0.8483, 0.7978, 0.8576, 0.8598, and 0.8216, respectively. The average AUC was 0.8370 ± 0.0238, and the MCC values were 0.41, 0.35, 0.40, 0.41, and 0.41, respectively. The calibration plot and DCA in the testing group also showed good performance. The AUC values and confidence intervals of the 1000-iteration model were: logistic (AUC, min confidence interval [CI]-max confidence interval [CI] 0.848, 0.804-0.903), RF (AUC, minCI-maxCI 0.797, 0.734-0.857), MLP (AUC, minCI-maxCI 0.858, 0.812-0.902), SVM (AUC, minCI-maxCI 0.859, 0.819-0.910), and XGB (AUC, minCI-maxCI 0.821, 0.764-0.894). The model performed well. Finally, according to SHAP, among all five models, HB played the most important role in model prediction and interpretation. The five models we developed all performed well in predicting the likelihood of blood transfusion after hip fracture surgery. Therefore, we believed that the prediction model based on machine learning had great application prospects in clinical practice, which could help clinicians better predict the risk of blood transfusion after hip fracture surgery.

Savannah grazing lawns are a key food resource for large herbivores such as blue wildebeest (Connochaetes taurinus), hippopotamus (Hippopotamus amphibius) and white rhino (Ceratotherium simum), and impact herbivore densities, movement and recruitment rates. They also exert a strong influence on fire behaviour including frequency, intensity and spread. Thus, variation in grazing lawn cover can have a profound impact on broader savannah ecosystem dynamics. However, knowledge of their present cover and distribution is limited. Importantly, we lack a robust, broad-scale approach for detecting and monitoring grazing lawns, which is critical to enhancing understanding of the ecology of these vital grassland systems. We selected two sites in the Lower Sabie and Satara regions of Kruger National Park, South Africa with mesic and semiarid conditions, respectively. Using spectral and texture features derived from WorldView-3 imagery, we (i) parameterised and assessed the quality of Random Forest (RF), Support Vector Machines (SVM), Classification and Regression Trees (CART) and Multilayer Perceptron (MLP) models for general discrimination of plant functional types (PFTs) within a sub-area of the Lower Sabie landscape, and (ii) compared model performance for probabilistic mapping of grazing lawns in the broader Lower Sabie and Satara landscapes. Further, we used spatial metrics to analyse spatial patterns in grazing lawn distribution in both landscapes along a gradient of distance from waterbodies. All machine learning models achieved high F-scores (F1) and overall accuracy (OA) scores in general savannah PFTs classification, with RF (F1 = 95.73±0.004%, OA = 94.16±0.004%), SVM (F1 = 95.64±0.002%, OA = 94.02±0.002%) and MLP (F1 = 95.71±0.003%, OA = 94.27±0.003%) forming a cluster of the better performing models and marginally outperforming CART (F1 = 92.74±0.006%, OA = 90.93±0.003%). Grazing lawn detection accuracy followed a similar trend within the Lower Sabie landscape, with RF, SVM, MLP and CART achieving F-scores of 0.89, 0.93, 0.94 and 0.81, respectively. Transferring models to the Satara landscape however resulted in relatively lower but high grazing lawn detection accuracies across models (RF = 0.87, SVM = 0.88, MLP = 0.85 and CART = 0.75). Results from spatial pattern analysis revealed a relatively higher proportion of grazing lawn cover under semiarid savannah conditions (Satara) compared to the mesic savannah landscape (Lower Sabie). Additionally, the results show strong negative correlation between grazing lawn spatial structure (fractional cover, patch size and connectivity) and distance from waterbodies, with larger and contiguous grazing lawn patches occurring in close proximity to waterbodies in both landscapes. The proposed machine learning approach provides a novel and robust workflow for accurate and consistent landscape-scale monitoring of grazing lawns, while our findings and research outputs provide timely information critical for understanding habitat heterogeneity in southern African savannahs.