Flood Mapping with Convolutional Neural Networks Using Spatio-Contextual Pixel Information

Arterio Venous Vascular Access (AVA) stenosis is a common complication in hemodialysis patients. Clinically, AVA stenosis happens when the cross-sectional area is reduced to less than 50% of the normal area. In order to highlight the correlation between the power spectral features of AVA Phonoangiography signals (PCG) and degree of stenosis (DOS), In-vitro BioPhysical Simulation Model (BPSM) is used to control individual conditions. Previous studies have pointed out that the features of PCG can be used to detect stenosis, but there were differences in the specific frequency band range. In this study, a method for extracting the power spectral features of PCG based on MUltiple SIgnal Classification power spectrum estimation algorithm (MUSIC) is proposed. This method has a high resolution for the high-frequency low-energy sound caused by stenosis. Using the proposed method, a strong correlation is found between the frequency peak near 820 Hz (820 ±70 Hz) and the AVA stenosis. Based on the above feature extraction method, a support vector machine (SVM) classification model is trained on data obtained on the BPSM. Finally, using MUSIC features extraction model and SVM classification model, the correct classification rate on BPSM data is 96.4%, and the SVM model is validated on 19 clinical measured data, the accuracy is 84.2%.

In breast cancer diagnosis and treatment, non-invasive prediction of axillary lymph node (ALN) metastasis can help avoid complications related to sentinel lymph node biopsy. This study aims to develop and evaluate machine learning models using radiomics features extracted from diffusion-weighted whole-body imaging with background signal suppression (DWIBS) examination for predicting the ALN status. A total of 100 patients with histologically proven, invasive, clinically N0 breast cancer who underwent DWIBS examination consisting of short tau inversion recovery (STIR) and DWIBS sequences before surgery were enrolled. Radiomic features were calculated using segmented primary lesions in DWIBS and STIR sequences and were divided into training (n = 75) and test (n = 25) datasets based on the examination date. Using the training dataset, optimal feature selection was performed using the least absolute shrinkage and selection operator algorithm, and the logistic regression model and support vector machine (SVM) classifier model were constructed with DWIBS, STIR, or a combination of DWIBS and STIR sequences to predict ALN status. Receiver operating characteristic curves were used to assess the prediction performance of radiomics models. For the test dataset, the logistic regression model using DWIBS, STIR, and a combination of both sequences yielded an area under the curve (AUC) of 0.765 (95% confidence interval: 0.548-0.982), 0.801 (0.597-1.000), and 0.779 (0.567-0.992), respectively, whereas the SVM classifier model using DWIBS, STIR, and a combination of both sequences yielded an AUC of 0.765 (0.548-0.982), 0.757 (0.538-0.977), and 0.779 (0.567-0.992), respectively. Use of machine learning models incorporating with the quantitative radiomic features derived from the DWIBS and STIR sequences can potentially predict ALN status.

Seed-per-pod estimation for plant breeding using deep learning

Classification of suspect powders, by using laser‐induced breakdown spectroscopy (LIBS) spectra, to determine if they could contain Bacillus anthracis spores is difficult because of the variability in their composition and the variability typically associated with LIBS analysis. A method that builds a support vector machine classification model for such spectra relying on the known elemental composition of the Bacillus spores was developed. A wavelet transformation was incorporated in this method to allow for possible thresholding or standardization, then a linear model technique using the known elemental structure of the spores was incorporated for dimension reduction, and a support vector machine approach was employed for the final classification of the substance. The method was applied to real data produced from an LIBS device. Several methods used to test the predictive performance of the classification model revealed promising results. Published 2012. This article is a US Government work and is in the public domain in the USA.

Recognition of necrotic regions in MRI images of chronic spinal cord injury based on superpixel

An automated non-destructive prediction of peroxide value and free fatty acid level in mixed nut samples

I335 the training of Support Vector Machine (SVM) classification model, some problems such as over-fitting, under-fitting, slow training speed and prediction affected by penalty and kernel function parameters are exposed. A fault classification method of rolling bearing based on Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and Grey Wolf Optimization (GWO) -SVM is proposed. Firstly, the acquire signals are adaptively decomposed into a plurality of Intrinsic Mode Function(IMF) component by CEEMDAN, and the energy entropy of each IMF component is extracted to form a set of high dimensional feature vector. Then, the processed feature vectors are introduced into the diagnosis network of GWO-SVM algorithm to build a fault classification model of motor rolling bearing. The classification results show that the CEEMDAN and GWO-SVM fault classification network of motor rolling bearing has higher accuracy and shorter diagnosis time than Particle Swarm Optimization (PSO) -SVM and Cross Validation (CV) -SVM.

Identification of SARS-CoV-2 viral entry inhibitors using machine learning and cell-based pseudotyped particle assay

In the realm of machine learning, classification models are important for identifying patterns and grouping data. Support Vector Machine (SVM) and Robust SVM are two types of models that are often used. SVM works by finding an optimal hyperplane to separate data classes, while Robust SVM is designed to deal with uncertainty and noise in the data, making it more resistant to outliers. However, SVM has limitations in dealing with class imbalance and outliers in the dataset. Class imbalance makes the model tend to predict the majority class, and outliers can interfere with model formation. This research compares the performance of SVM and Robust SVM on normal, unbalanced and outlier datasets. The software uses Python and Scikit-learn for implementation and comparison of the two models. Key features include automatic data preprocessing, model training, and evaluation with metrics such as accuracy, precision, recall, and F1 score. The results show that Robust SVM is superior in accuracy on normal datasets and is very effective in dealing with class imbalance, achieving a maximum accuracy of 100%. On datasets with outliers, Robust SVM maintains stable accuracy, demonstrating its robustness to outliers. This research contributes to correspondence management by providing more reliable classification models, improving data processing accuracy, and supporting more informed decision making in software development

This article aims at finding the risk factors for hepatitis B virus (HBV) reactivation after the precise radiotherapy in patients with primary liver cancer (PLC). We use sequential forward selection and sequential backward selection to extract features which would be combined into an optimal feature subset, and then establish Bayesian and support vector machine (SVM) classification model. We use sequential forward selection to select the key features and find that the AFP, HBV DNA levels, outer margin of radiotherapy, times of radiotherapy, split method were the risk factors of HBV reactivation. The accuracy of the key features with Bayesian classifier reached to 80.65% by using 5 fold cross validation and with SVM reached to 79.84% by using 10 fold cross validation. Besides, we use sequential backward selection find that KPS score, HBV DNA level, outer margin of radiotherapy, tumor stage TNM, equivalent biometric were the risk factors of HBV reactivation, meanwhile, the accuracy of Bayesian classifier can be reached to 85.77% by using fold cross validation and with SVM classifier can be reached to 87.31% by using 10 fold cross validation. The accuracy of original feature with Bayesian classifier reached to 71.42% by using 10 fold cross validation and with SVM classifier reached to 78.10% by using 5 fold cross validation. The experimental results showed that the key feature subset has a better classification performance than the initial feature set clearly.

Current standards for identifying the characteristics of Qingpi (QP) primarily depend on traditional empirical methods, which are both time-consuming and labor-intensive. Therefore, there is significant potential for developing a rapid and nondestructive identification method. This work aims to realize fast and effective identification of QP and its processed products, to provide a basis for quality assurance and monitoring of QP and its processed products, and to provide reference for guiding the practical work of traditional Chinese medicine (TCM) identification. In this study, we employed three advanced technologies-electronic eye (E-eye), fast gas chromatography electronic nose (Fast GC e-nose), and Fourier transform near-infrared (FT-NIR) spectroscopy-to analyze the color, odor, and absorbance of QP samples. The E-eye digitized the appearance color of various QP samples, and subsequent discriminant analysis, combined with statistical methods, confirmed the feasibility of the discriminant function obtained. Additionally, the Fast GC e-nose provided valuable odor information, identifying 18 distinct odor components. Based on variable importance in projection (VIP) analysis, four components were hypothesized as potential odor markers for distinguishing QP raw products (SQP) from vinegar processed products (CQP). According to the accuracy of the support vector machine (SVM) classification model, the NIR preprocessing method is screened. Alongside SVM, three classification models are chosen for simultaneous evaluation and verification. Notably, the test set recognition rate for all four classification models reached 100%. E-eye, Fast GC e-nose, and NIR technology enable rapid, nondestructive identification and preliminary quality evaluation of QP products.

Cancer is one of the leading causes of death, and breast cancer is the second leading cause of cancer death in women. One method to realize the level of malignancy of breast cancer from an early age is by classifying the cancer malignancy using data mining. One of the widely used data mining methods with a good level of accuracy is the Support Vector Machine (SVM) and K-Nearest Neighbors (KNN). Evaluation techniques of percentage split and cross-validation were used to evaluate and compare the SVM and KNN classification models. The result was that the accuracy level of the SVM classification method was better than the KNN classification method when using the cross-validation technique, which is 95,7081%. Meanwhile, the KNN classification method was better than the SVM classification method when using the percentage split technique, which is 95,4220%. From the comparison results, it can be seen that the KNN and SVM methods work well in the classification of breast cancer.

This study presents the development of artificial neural network (ANN) and support vector machine (SVM) classification models for predicting the safety conditions of water in distribution pipes. The study was based on 504 monthly records of water quality parameters; pH, turbidity, color and bacteria counts taken from nine different locations across the water distribution network in the city of Alesund, Norway. The models predicted the safety conditions of the water samples in the pipes with 98% accuracy and 94% respectively during testing. The high accuracy achieved in the model results indicate that contamination events in distribution systems that result in unsafe values of the water quality parameters can be detected using these classification models. This can provide water utility managers with real time information about the safety conditions of treated water at different locations of distribution pipes before water reaches consumers.

Simple SummaryMalignant melanomas (MMs) with aypical clinical presentation constitute a diagnostic pitfall, and false negatives carry the risk of a diagnostic delay and improper disease management. Among the most common, challenging presentation forms of MMs are those that clinically resemble seborrheic keratosis (SK). On the other hand, SK may mimic melanoma, producing ‘false positive overdiagnosis’ and leading to needless excisions. The evolving efficiency of deep learning algorithms in image recognition and the availability of large image databases have accelerated the development of advanced computer-aided systems for melanoma detection. In the present study, we used image data from the International Skin Image Collaboration archive to explore the capacity of deep knowledge transfer in the challenging diagnostic task of the atypical skin tumors of MM and SK.Malignant melanomas resembling seborrheic keratosis (SK-like MMs) are atypical, challenging to diagnose melanoma cases that carry the risk of delayed diagnosis and inadequate treatment. On the other hand, SK may mimic melanoma, producing a ‘false positive’ with unnecessary lesion excisions. The present study proposes a computer-based approach using dermoscopy images for the characterization of SΚ-like MMs. Dermoscopic images were retrieved from the International Skin Imaging Collaboration archive. Exploiting image embeddings from pretrained convolutional network VGG16, we trained a support vector machine (SVM) classification model on a data set of 667 images. SVM optimal hyperparameter selection was carried out using the Bayesian optimization method. The classifier was tested on an independent data set of 311 images with atypical appearance: MMs had an absence of pigmented network and had an existence of milia-like cysts. SK lacked milia-like cysts and had a pigmented network. Atypical MMs were characterized with a sensitivity and specificity of 78.6% and 84.5%, respectively. The advent of deep learning in image recognition has attracted the interest of computer science towards improved skin lesion diagnosis. Open-source, public access archives of skin images empower further the implementation and validation of computer-based systems that might contribute significantly to complex clinical diagnostic problems such as the characterization of SK-like MMs.

The purpose of this experiment is to utilize attenuated total reflection (ATR) Fourier transform infrared (FTIR) spectroscopy for the discrimination of different types of hair, as numerous studies have substantiated its efficacy in substance classification. In this study, ATR-FTIR spectroscopy was employed to analyze scalp hair, pubic hair, and armpit hair from human subjects. Additionally, a machine learning model was integrated to differentiate between hairs originating from distinct body regions. Because of the limited sampling conditions, we only chose samples from Chinese people who have been living in Shanghai and the surrounding areas for a long time to conduct the experiment. We developed partial least squares discriminant analysis (PLS-DA), random forest (RF), and support vector machine (SVM) classification models and compared their performance in identification. The results show that the SVM model has the best identification results with 90.37% accuracy, 90.37% recall, and 90.38% precision. This preliminary study suggests that ATR-FTIR spectroscopy combined with SVM may be an effective and promising aid in assisting the identification of hair in different parts of the human body. This method is non-destructive, fast, and accurate, and does not require a sample preparation process, which makes it promising in the field of forensic science. Also, we found that the main substance differences that contributed to the good distinction between hairs were expressed in amide I, followed by amide III and C-H deformation.