Support Vector Machine Classifier Research Articles

Predicting the Response of Patients Treated with 177Lu-DOTATATE Using Single-photon Emission Computed Tomography–Computed Tomography Image-based Radiomics and Clinical Features

Abstract Background: In this study, we want to evaluate the response to Lutetium-177 (177Lu)-DOTATATE treatment in patients with neuroendocrine tumors (NETs) using single-photon emission computed tomography (SPECT) and computed tomography (CT), based on image-based radiomics and clinical features. Methods: The total volume of tumor areas was segmented into 61 SPECT and 41 SPECT-CT images from 22 patients with NETs. A total of 871 radiomics and clinical features were extracted from the SPECT and SPECT-CT images. Subsequently, a feature reduction method called maximum relevance minimum redundancy (mRMR) was used to select the best combination of features. These selected features were modeled using a decision tree (DT), random forest (RF), K-nearest neighbor (KNN), and support vector machine (SVM) classifiers to predict the treatment response in patients. For the SPECT and SPECT-CT images, ten and eight features, respectively, were selected using the mRMR algorithm. Results: The results revealed that the RF classifier with feature selection algorithms through mRMR had the highest classification accuracies of 64% and 83% for the SPECT and SPECT-CT images, respectively. The accuracy of the classifications of DT, KNN, and SVM for SPECT-CT images is 79%, 74%, and 67%, respectively. The poor accuracy obtained from different classifications in SPECT images (≈64%) showed that these images are not suitable for predicting treatment response. Conclusions: Modeling the selected features of SPECT-CT images based on their anatomy and the presence of extensive gray levels makes it possible to predict responses to the treatment of 177Lu-DOTATATE for patients with NETs.

Colorectal cancer histopathology image analysis: A comparative study of prognostic values of automatically extracted morphometric nuclear features in multispectral and red-blue-green imagery.

Multispectral imaging (MSI) has been utilized to predict the prognosis of colorectal cancer (CRC) patients, however, our understanding of the prognostic value of nuclear morphological parameters of bright-field MSI in CRC is still limited. This study was designed to compare the efficiency of MSI and standard red-green-blue (RGB) images in predicting the prognosis of CRC. We compared the efficiency of MS and conventional RGB images on the quantitative assessment of hematoxylin-eosin (HE) stained histopathology images. A pipeline was developed using a pixel-wise support vector machine (SVM) classifier for gland-stroma segmentation, and a marker-controlled watershed algorithm was used for nuclei segmentation. The correlation between extracted morphological parameters and the five-year disease-free survival (5-DFS) was analyzed. Forty-seven nuclear morphological parameters were extracted in total. Based on Kaplan-Meier analysis, eight features derived from MS images and seven featured derived from RGB images were significantly associated with 5-DFS, respectively. Compared with RGB images, MSI showed higher accuracy, precision, and Dice index in nuclei segmentation. Multivariate analysis indicated that both integrated parameters 1 (factors negatively correlated with CRC prognosis including nuclear number, circularity, eccentricity, major axis length) and 2 (factors positively correlated with CRC prognosis including nuclear average area, area perimeter, total area/total perimeter ratio, average area/perimeter ratio) in MS images were independent prognostic factors of 5-DFS, in contrast with only integrated parameter 1 (P<0.001) in RGB images. More importantly, the quantification of HE-stained MS images displayed higher accuracy in predicting 5-DFS compared with RGB images (76.9% vs 70.9%). Quantitative evaluation of HE-stained MS images could yield more information and better predictive performance for CRC prognosis than conventional RGB images, thereby contributing to precision oncology.

Confidential outsourced support vector machine learning based on well-separated structure

Support Vector Machine (SVM) has revolutionized various domains and achieved remarkable successes. This progress relies on subtle algorithms and more on large training samples. However, the massive data collection introduces security concerns. To facilitate secure integration of data efficiently for building an accurate SVM classifier, we present a non-interactive protocol for privacy-preserving SVM, named NPSVMT. Specifically, we define a new well-separated structure for computing gradients that can decouple the fusion matter between user data and model parameters, allowing data providers to outsource the collaborative learning task to the cloud. As a result, NPSVMT is capable of removing the multiple communications and eliminating the straggler’s effect (waiting for the last), thereby going beyond those developed with interactive methods, e.g., federated learning. To further decrease the data traffic, we introduce a high-efficient coding method to compress and parse training data. In addition, unlike outsourced schemes based on homomorphic encryption or secret sharing, NPSVMT exploits functional encryption to maintain the data confidentiality, achieving dropout-tolerant secure aggregation. The implementations verify that NPSVMT is faster by orders of magnitude than the existing privacy-preserving SVM schemes on benchmark datasets.

Multi-view support vector machine classifier via [formula omitted] soft-margin loss with structural information

Multi-view learning seeks to leverage the advantages of various views to complement each other and make full use of the latent information in the data. Nevertheless, effectively exploring and utilizing common and complementary information across diverse views remains challenging. In this paper, we propose two multi-view classifiers: multi-view support vector machine via L0/1 soft-margin loss (MvL0/1-SVM) and structural MvL0/1-SVM (MvSL0/1-SVM). The key difference between them is that MvSL0/1-SVM additionally fuses structural information, which simultaneously satisfies the consensus and complementarity principles. Despite the discrete nature inherent in the L0/1 soft-margin loss, we successfully establish the optimality theory for MvSL0/1-SVM. This includes demonstrating the existence of optimal solutions and elucidating their relationships with P-stationary points. Drawing inspiration from the P-stationary point optimality condition, we design and integrate a working set strategy into the proximal alternating direction method of multipliers. This integration significantly enhances the overall computational speed and diminishes the number of support vectors. Last but not least, numerical experiments show that our suggested models perform exceptionally well and have faster computational speed, affirming the rationality and effectiveness of our methods.

Fusion of Wavelet Features and Gabor Features for SVM-based Iris Verification

Iris verification now become increasingly prominent in biometric-based person verification systems. It has gained a significant role in biometric systems due to its stability, high uniqueness, contactless and non-invasive properties. Iris has more inherent distinctive features than other biometrics. Feature extraction of iris plays a crucial role in this system for accurate person verification. Using the feature extraction process, unique iris features like textural patterns, crypts, and furrows of iris are extracted. In our study, we did a fusion of Discrete Wavelet Transform (DWT) features with multiple wavelet bases (db4, haar, coif3, and sym4) and Gabor features, which contain a good amount of textural and localized information. Fusion here indicates the concatenation of the extracted features using the above techniques. In this work, we studied this method on the full iris only so that a maximum number of features can be extracted. This combined approach yielded a significant 112 number of features. The extracted features are then verified using a support vector machine (SVM) classifier based on radial basis function (RBF) kernel with training vs testing split ratios of 8:2, 6:4, 4:6 and 2:8. In this study, we achieved the highest overall verification accuracy of 95.9% with training vs testing split ratio of 8:2. For other training vs testing split ratios of 6:4, 4:6 and 2:8 we achieved overall verification accuracies of 91.4%, 93.2% and 91.2% respectively. We got an overall verification accuracy of 92.9%, considering training vs testing ratios of 8:2, 6:4, 4:6 and 2:8.

Analyzing Resampling Techniques for Addressing the Class Imbalance in NIDS using SVM with Random Forest Feature Selection

The purpose of Network Intrusion Detection Systems (NIDS) is to ensure and protect computer networks from harmful actions. A major concern in NIDS development is the class imbalance problem, i.e., normal traffic dominates the communication data plane more than intrusion attempts. Such a state of affairs can pose certain hazards to the effectiveness of detection algorithms, including those useful for detecting less frequent but still highly dangerous intrusions. This paper aims to utilize resampling techniques to tackle this problem of class imbalance in NIDS using a Support Vector Machine (SVM) classifier alongside utilizing features selected by Random Forest to improve the feature subset selection process. The analysis highlights the combativeness of each sampling method, offering insights into their efficiency and practicality for real-world applications. Four resampling techniques are analyzed. Such techniques include Synthetic Minority Over-sampling Technique (SMOTE), Random Under-sampling (RUS), Random Over-sampling (ROS) and SMOTE with two different combinations i.e., RUS SMOTE and RUS ROS. Feature selection was done using Random Forest, which was improved by Bayesian methods to create subsets of features with feature rankings determined by Cumulative Feature Importance Score (CFIS). The CIDDS-2017 dataset is used for the performance evaluation, and the metrics used include accuracy, precision, recall, F-measure and CPU time. The algorithm that performs best overall in the CFIS feature subsets is SMOTE, and the features that give the best result are selected at the 90% level with 25 features. This subset accomplishes a relative accuracy enhancement of 0.08% than the other approaches. The RUS+ROS technique is also fine but somehow slower than SMOTE. On the other hand, RUS+SMOTE shows relatively poor results although it consumes less time in terms of computational time compared to other methods, giving about 50% of the performance shown by the other methods. This paper's novelty is adapting the RUS method as a standalone test for screening new and potentially contaminated datasets. The standalone RUS method is more efficient in terms of computations; the algorithm returned the best result of 98.13% accuracy at 85% at the CFIS level of 34 features with a computation time of 137.812 s. It is also noted that SMOTE is considered to be proficient among all resampling techniques used for handling the problem of class imbalance in NIDS, vice 90% CFIS feature subset. Future research directions could include using these techniques in different data sets and other machine learning and deep learning methods together with ROC curve analysis to provide useful pointers to NIDS designers on how to select the right data mining tools and strategies for their projects.

Architecture of an automated program complex based on a multiple kernel svm classifier for analyzing malicious executable files

Subject matter. This article presents the development and architecture of an automated program complex designed to identify and analyze malicious executable files using a classifier based on a multiple kernel support vector machine (SVM). Goal. The aim of the work is to create an automated system that enhances the accuracy and efficiency of malware detection by combining static and dynamic analysis into a single framework capable of processing large volumes of data with optimal time expenditure. Tasks. To achieve this goal, tasks were carried out that included developing a program complex that automates the collection of static and dynamic data from executable files using tools like IDA Pro, IDAPython, and Drakvuf; integrating a multiple kernel SVM classifier to analyze the collected heterogeneous data; validating the system's effectiveness based on a substantial dataset containing 1,389 executable samples; and demonstrating the system's scalability and practical applicability in real-world conditions. Methods. The methods involved a hybrid approach that combines static analysis – extracting byte code, disassembled instructions, and control flow graphs using IDA Pro and IDAPython – with dynamic analysis, which entails monitoring real-time behavior using Drakvuf. The multiple kernel SVM classifier integrates different data representations using various kernels, allowing for both linear and nonlinear relationships to be considered in the classification process. Results. The results of the study show that the system achieves a high level of accuracy and completeness, as evidenced by key performance metrics such as an F-score of 0.93 and ROC AUC and PR AUC values. The automated program complex reduces the analysis time of a single file from an average of 11 minutes to approximately 5 minutes, effectively doubling the throughput compared to previous methods. This significant reduction in processing time is critically important for deployment in environments where rapid and accurate malware detection is necessary. Furthermore, the system's scalability allows for efficient processing of large data volumes, making it suitable for real-world applications. Conclusions. In conclusion, the automated program complex developed in this study demonstrates significant improvements in the accuracy and efficiency of malware detection. By integrating multiple kernel SVM classification with static and dynamic analysis, the system shows potential for real-time malware detection and analysis. Its scalability and practical applicability indicate that it could become an important tool in combating modern cyber threats, providing organizations with an effective means to enhance their cybersecurity.

Fractal-based supervised approach for dimensionality reduction of hyperspectral images

Dimensionality reduction is one of the most challenging and crucial issues apart from data mining, security, and scalability, which have retained much traction due to the ever-growing need to analyze the large volumes of data generated daily. Fractal Dimension (FD) has been successfully used to characterize data sets and has found relevant applications in dimension reduction. This paper presents an application of the FD Reduction (FDR) Algorithm on geospatial hyperspectral data, examining its usefulness for data sets with a relatively high embedding dimension. We examine the algorithm at two levels. First is the conventional FDR approach (unsupervised) at the image level. Alternatively, we propose a pixel-level supervised approach for band reduction based on time-series complexity analysis. Techniques for determining an optimal intrinsic dimension for the dataset using these two techniques are examined. We also develop a parallel GPU-based implementation for the unsupervised image-level FDR algorithm, reducing the run-time by nearly 10 times. Furthermore, both approaches use a support vector machine classifier to compare the classification performance of the original and reduced image obtained.

An Optimal Feature Selection Method for Human Activity Recognition Using Multimodal Sensory Data

Recently, the research community has taken great interest in human activity recognition (HAR) due to its wide range of applications in different fields of life, including medicine, security, and gaming. The use of sensory data for HAR systems is most common because the sensory data are collected from a person’s wearable device sensors, thus overcoming the privacy issues being faced in data collection through video cameras. Numerous systems have been proposed to recognize some common activities of daily living (ADLs) using different machine learning, image processing, and deep learning techniques. However, the existing techniques are computationally expensive, limited to recognizing short-term activities, or require large datasets for training purposes. Since an ADL is made up of a sequence of smaller actions, recognizing them directly from raw sensory data is challenging. In this paper, we present a computationally efficient two-level hierarchical framework for recognizing long-term (composite) activities, which does not require a very large dataset for training purposes. First, the short-term (atomic) activities are recognized from raw sensory data, and the probabilistic atomic score of each atomic activity is calculated relative to the composite activities. In the second step, the optimal features are selected based on atomic scores for each composite activity and passed to the two classification algorithms: random forest (RF) and support vector machine (SVM) due to their well-documented effectiveness for human activity recognition. The proposed method was evaluated on the publicly available CogAge dataset that contains 890 instances of 7 composite and 9700 instances of 61 atomic activities. The data were collected from eight sensors of three wearable devices: a smartphone, a smartwatch, and smart glasses. The proposed method achieved the accuracy of 96.61% and 94.1% by random forest and SVM classifiers, respectively, which shows a remarkable increase in the classification accuracy of existing HAR systems for this dataset.

Ensemble Fusion Models Using Various Strategies and Machine Learning for EEG Classification.

Electroencephalography (EEG) helps to assess the electrical activities of the brain so that the neuronal activities of the brain are captured effectively. EEG is used to analyze many neurological disorders, as it serves as a low-cost equipment. To diagnose and treat every neurological disorder, lengthy EEG signals are needed, and different machine learning and deep learning techniques have been developed so that the EEG signals could be classified automatically. In this work, five ensemble models are proposed for EEG signal classification, and the main neurological disorder analyzed in this paper is epilepsy. The first proposed ensemble technique utilizes an equidistant assessment and ranking determination mode with the proposed Enhance the Sum of Connection and Distance (ESCD)-based feature selection technique for the classification of EEG signals; the second proposed ensemble technique utilizes the concept of Infinite Independent Component Analysis (I-ICA) and multiple classifiers with majority voting concept; the third proposed ensemble technique utilizes the concept of Genetic Algorithm (GA)-based feature selection technique and bagging Support Vector Machine (SVM)-based classification model. The fourth proposed ensemble technique utilizes the concept of Hilbert Huang Transform (HHT) and multiple classifiers with GA-based multiparameter optimization, and the fifth proposed ensemble technique utilizes the concept of Factor analysis with Ensemble layer K nearest neighbor (KNN) classifier. The best results are obtained when the Ensemble hybrid model using the equidistant assessment and ranking determination method with the proposed ESCD-based feature selection technique and Support Vector Machine (SVM) classifier is utilized, achieving a classification accuracy of 89.98%.

Comparative analysis of machine learning techniques for detecting potability of water

This research aims to evaluate the effectiveness of machine learning algorithms in determining the potability of water. In the study, a total of 3276 water samples were analyzed for 10 different features that determine the potability of water. Besides that, the study's consideration is to evaluate the impact of trimming, IQR, and percentile methods on the performance of machine learning algorithms. The models were built using nine different classification algorithms (Logistic Regression, Decision Trees, Random Forest, XGBoost, Naive Bayes, K-Nearest Neighbors, Support Vector Machine, AdaBoost, and Bagging Classifier). According to the results, filling the missing data with the population mean and handling outliers with Trimming and IQR methods improved the performance of the models. Random Forest and Decision Tree algorithms were the most accurate in determining the potability of water. The findings of this research are of high importance to sustainable water resource management and serve as a crucial input for the decision-making process on the quality of water. The study also offers an example for researchers working on datasets that contain missing values and outliers.

Identifying defects and varieties of Malting Barley Kernels

This study introduces a comprehensive approach for classifying individual malting barley kernels, involving dual-sided kernel imaging, a specifically designed image processing algorithm, an optimized deep neural network architecture, and a mechanical sorting system. The proposed method achieves precise classification into multiple classes, aligning with quality standards for malting material assessment. Throughout the study, various image analysis techniques were assessed, including traditional feature engineering, established transfer learning deep neural network architectures, and our custom-designed convolutional neural network tailored for barley kernel image analysis. Comparative analysis underscores the superior performance of our network model. The study reveals that our proposed deep learning network achieves a 94% accuracy in classifying barley kernel defects and varieties, outperforming well-established transfer learning models to complex architectures that attain 93% accuracy. Additionally, it surpasses the traditional machine learning approach involving feature extraction and support vector machine classifiers, which achieve accuracy below 90% in detecting defective kernels and below 70% in varietal classification. However, we also noted the traditional approach’s advantage in morphological feature recognition. This observation guides new research toward integrating morphological feature extraction techniques with modern convolutional networks. This paper presents a deep neural network designed specifically for the analysis of cereal kernel images in two applications: defect and variety classification. It emphasizes the importance of standardizing kernel orientation and merging images from both sides of the kernel, and introduces a device for image acquisition that fulfills this need.

RanBALL: An Ensemble Random Projection Model for Identifying Subtypes of B-cell Acute Lymphoblastic Leukemia.

As the most common pediatric malignancy, B-cell acute lymphoblastic leukemia (B-ALL) has multiple distinct subtypes characterized by recurrent and sporadic somatic and germline genetic alterations. Identification of B-ALL subtypes can facilitate risk stratification and enable tailored therapeutic approaches. Existing methods for B-ALL subtyping primarily depend on immunophenotypic, cytogenetic and genomic analyses, which would be costly, complicated, and laborious in clinical practice applications. To overcome these challenges, we present RanBALL (an Ensemble Ran dom Projection-Based Model for Identifying B -Cell A cute L ymphoblastic L eukemia Subtypes), an accurate and cost-effective model for B-ALL subtype identification based on transcriptomic profiling only. RanBALL leverages random projection (RP) to construct an ensemble of dimension-reduced multi-class support vector machine (SVM) classifiers for B-ALL subtyping. Results based on 100 times 5-fold cross validation tests for >1700 B-ALL patients demonstrated that the proposed model achieved an accuracy of 93.35%, indicating promising prediction capabilities of RanBALL for B-ALL subtyping. The high accuracies of RanBALL suggested that our model could effectively capture underlying patterns of transcriptomic profiling for accurate B-ALL subtype identification. We believe RanBALL will facilitate the discovery of B-ALL subtype-specific marker genes and therapeutic targets, and eventually have consequential positive impacts on downstream risk stratification and tailored treatment design.

Validation of a Textile-Based Wearable Measuring Electrocardiogram and Breathing Frequency for Sleep Apnea Monitoring

Sleep apnea (SA) is a prevalent disorder characterized by recurrent events of nocturnal apnea. Polysomnography (PSG) represents the gold standard for SA diagnosis. This laboratory-based procedure is complex and costly, and less cumbersome wearable devices have been proposed for SA detection and monitoring. A novel textile multi-sensor monitoring belt recording electrocardiogram (ECG) and breathing frequency (BF) measured by thorax excursion was developed and tested in a sleep laboratory for validation purposes. The aim of the current study was to evaluate the diagnostic performance of ECG-derived heart rate variability and BF-derived breathing rate variability and their combination for the detection of sleep apnea in a population of patients with a suspicion of SA. Fifty-one patients with a suspicion of SA were recruited in the sleep laboratory of the Cantonal Hospital St. Gallen. Patients were equipped with the monitoring belt and underwent a single overnight laboratory-based PSG. In addition, some patients further tested the monitoring belt at home. The ECG and BF signals from the belt were compared to PSG signals using the Bland-Altman methodology. Heart rate and breathing rate variability analyses were performed. Features derived from these analyses were used to build a support vector machine (SVM) classifier for the prediction of SA severity. Model performance was assessed using receiver operating characteristics (ROC) curves. Patients included 35 males and 16 females with a median age of 49 years (range: 21 to 65) and a median apnea-hypopnea index (AHI) of 33 (IQR: 16 to 58). Belt-derived data provided ECG and BF signals with a low bias and in good agreement with PSG-derived signals. The combined ECG and BF signals improved the classification accuracy for SA (area under the ROC curve: 0.98; sensitivity and specificity greater than 90%) compared to single parameter classification based on either ECG or BF alone. This novel wearable device combining ECG and BF provided accurate signals in good agreement with the gold standard PSG. Due to its unobtrusive nature, it is potentially interesting for multi-night assessments and home-based patient follow-up.

Sensor-Embedded Tissue Phantom for Magnetic Resonance Elastography Mechanical Failure Testing

Abstract Magnetic Resonance Elastography (MRE) is an imaging technique capable of quantifying the stiffness of in vivo tissue by applying and imaging shear waves produced by an MRE actuator. Poor image acquisition may result from the MRE procedure if there is insufficient contact between the MRE actuator and the patient. An experimental test setup outside of the clinic will aid in reducing the number of failed acquisitions by enabling the development of advanced actuators and actuator systems. This work presents the development and testing of a sensor-embedded tissue phantom setup paired with a support vector machine (SVM) classifier to automate the MRE actuator testing process. MRE actuation of soft tissue is simulated by utilizing a voice coil positioning stage that interfaces with a phantom. To capture the resulting vibrations, accelerometers are embedded inside the phantom. Subsequent characterization experiments verify the functionality of the developed phantoms to capture wave propagation. A secondary investigation was performed by utilizing the developed setup to collect acceleration measurements at varying contact distances. We provide an overview of feature analysis and selection to develop SVM models for contact detection. Multiple SVM models are reported, and the best-performing model displayed almost perfect validation (94.53%) and test (90.91%) accuracy. The pairing of sensor-embedded phantom with an SVM for detection demonstrates potential improvements to the MRE actuator developmental process by automatically assessing contact-related issues prior to clinical testing.

A Machine Learning-Based Wrapper Method for Feature Selection

This paper presents a two-stage feature selection scheme using machine learning techniques. In the first stage a wrapper method is adopted to select various combinations of subsets of features from the original dataset. The performance of the model is evaluated by three classifiers: K-Nearest Neighbor (KNN), Support Vector Machines (SVM), and Random Forest (RF). In the second and final stage, a sequential backward feature selection Method is applied. The proposed method is demonstrated on eighteen datasets and the average classification accuracy of eighteen datasets achieved is 89.81%, 87.55%, and 89.82% using the KNN, SVM, and RF classifiers, respectively with a maximum reduced size of the subset being ten only. Comparing the proposed method to eight other feature selection methods, the former achieves better classification accuracy in terms of selecting the most useful but a smaller number of features.

Cross-region feature fusion of global and local area for subtype classification prediction in cervical tumour

Background and objectivesTo develop a cross-regional feature fusion model for the classification of cervical tumour subtypes using non-invasive magnetic resonance imaging (MRI), we aimed to explore feature representation in both global and local areas, and to compare their effect on predictive performance. Method and materialsThis retrospective study included 100 patients with cervical cancer, approved by the Ethical Review Board Committee. Self-supervised learning-based global features were fused with local features for subtype classification modelling. Global features were extracted from the bottleneck of our 3D autoencoder network, while local features were derived based on a radiomics tool. Utilizing the global and local features, the classification model is based on machine learning algorithms to predict two subtypes with pathologically confirmed cervical squamous cell carcinoma and adenocarcinoma. Comparison performance was accessed using area under the curve (AUC), sensitivity, specificity, F1 score, and precision. ResultThe cross-regional feature fusion model showed the best performance (accuracy: 0.95 vs 0.65 in the fusion model and global model) by the support vector machines (SVM) classifier. Even when applied to axial slices with various classification methods, the fusion model consistently yields the best results. ConclusionOur approach preliminary evidence suggests that the fusion of global and local features provides a significant advantage in the clinical diagnosis of cervical cancer subtypes, warranting further investigation and potential application in cervical cancer diagnosis.

Feature extraction via 3-D homogeneous attribute decomposition for hyperspectral imagery classification

ABSTRACT Feature extraction is a core aspect in hyperspectral image classification, which can extract key information closely related to ground cover from complex scene, thus improving classification accuracy. Therefore, designing an effective feature extraction network is a hotspot and a challenge in the current research. In this paper, a feature extraction framework based on 3-D homogeneous attribute decomposition (3D-HAD) is proposed for HSI classification, which consists of the following key technologies. First, the principal component analysis algorithm is applied to the raw HSI to extract the principal components (PCs), and the raw HSI is clustered into many 3D superpixel blocks according to the first three PCs-based over-segmentation strategy. Then, a superpixel intrinsic attribute decomposition (SIAD) is designed to capture reflectance feature and suppress shading feature. Meanwhile, a metric entropy is introduced into the decomposition process to overcome the spectral-spatial weak assumption among pixels. Next, superpixel-guided recursive filtering is employed to preserve global details of HSI to enhance accuracy in HSI classification. Finally, the support vector machine classifier is used to obtain classification results of HSI. Experiments performed on several real hyperspectral data sets with limited training samples indicate that the proposed 3D-HAD method outperforms the classic, advance, and deep learning classification methods.

Assessing muscle invasion in bladder cancer via virtual biopsy: a study on quantitative parameters and classical radiomics features from dual-energy CT imaging

ObjectiveTo evaluate the prediction value of Dual-energy CT (DECT)-based quantitative parameters and radiomics model in preoperatively predicting muscle invasion in bladder cancer (BCa).Materials and methodsA retrospective study was performed on 126 patients with BCa who underwent DECT urography (DECTU) in our hospital. Patients were randomly divided into training and test cohorts with a ratio of 7:3. Quantitative parameters derived from DECTU were identified through univariate and multivariate logistic regression analysis to construct a DECT model. Radiomics features were extracted from the 40, 70, 100 keV and iodine-based material-decomposition (IMD) images in the venous phase to construct radiomics models from individual and combined images using a support vector machine classifier, and the optimal performing model was chosen as the final radiomics model. Subsequently, a fusion model combining the DECT parameters and the radiomics model was established. The diagnostic performances of all three models were evaluated through receiver operating characteristic (ROC) curves and the clinical usefulness was estimated using decision curve analysis (DCA).ResultsThe normalized iodine concentration (NIC) in DECT was an independent factor in diagnosing muscle invasion of BCa. The optimal multi-image radiomics model had predictive performance with an area-under-the-curve (AUC) of 0.867 in the test cohort, better than the AUC = 0.704 with NIC. The fusion model showed an increased level of performance, although the difference in AUC (0.893) was not statistically significant. Additionally, it demonstrated superior performance in DCA. For lesions smaller than 3 cm, the fusion model showed a high predictive capability, achieving an AUC value of 0.911. There was a slight improvement in model performance, although the difference was not statistically significant. This improvement was observed when comparing the AUC values of the DECT and radiomics models, which were 0.726 and 0.884, respectively.ConclusionThe proposed fusion model combing NIC and the optimal multi-image radiomics model in DECT showed good diagnostic capability in predicting muscle invasiveness of BCa.

Detection of hypoglycaemia in type 1 diabetes through breath volatile organic compound profiling using gas chromatography-ion mobility spectrometry.

To evaluate the relationship between breath volatile organic compounds (VOCs) and glycaemic states in individuals with type 1 diabetes (T1D), focusing on identifying specific VOCs as biomarkers for hypoglycaemia to offer a non-invasive diabetes-monitoring method. Ten individuals with T1D underwent induced hypoglycaemia in a clinical setting. Breath samples, collected every 10-15 minutes, were analysed using gas chromatography-ion mobility spectrometry (GC-IMS). Correlation analysis and machine learning models, including Partial Least Squares Discriminant Analysis (PLS-DA) and Support Vector Machine classifiers, were used to classify glycaemic states based on VOC profiles. Statistical analysis revealed moderate correlations between specific VOCs (e.g. isoprene, acetone) and venous blood glucose levels. Machine learning models showed high accuracy in classifying glycaemic states, with the best performance achieved by a two-class PLS-DA model showing an accuracy of 93%, sensitivity of 92% and specificity of 94%. Key biomarkers identified included isoprene, acetone, 2-butanone, methanol, ethanol, 2-propanol and 2-pentanone. This study shows the potential of breath VOCs to accurately classify glycaemic states in individuals with T1D. While key biomarkers such as isoprene, acetone and 2-butanone were identified, the analysis emphasizes the importance of using overall VOC patterns rather than individual compounds, which can be markers for multiple conditions. Machine learning models leveraging these patterns achieved high accuracy, sensitivity and specificity. These findings suggest that breath analysis using GC-IMS could be a viable non-invasive method for monitoring glycaemic states and managing diabetes.