Two-class Classification Research Articles

A non-linear optimization based robust attribute weighting model for the two-class classification problems.

This article aims to determine the coefficients that will reduce the in-class distance and increase the distance between the classes, collecting the data around the cluster centers with meta-heuristic optimization algorithms, thus increasing the classification performance. The proposed mathematical model is based on simple mathematical calculations, and this model is the fitness function of optimization algorithms. Compared to the methods in the literature, optimizing algorithms to obtain fast results is more accessible. Determining the weights by optimization provides more sensitive results than the dataset structure. In the study, the proposed model was used as the fitness function of the metaheuristic optimization algorithms to determine the weighting coefficients. In this context, four different structures were used to test the independence of the results obtained from the algorithm: the particle swarm algorithm (PSO), the bat algorithm (BAT), the gravitational search algorithm (GSA), and the flower pollination algorithm (FPA). As a result of these processes, a control group from unweighted attributes and four experimental groups from weighted attributes were obtained for each dataset. The classification performance of all datasets to which the weights obtained by the proposed method were applied increased. 100% accuracy rates were obtained in the Iris and Liver Disorders datasets used in the study. From synthetic datasets, from 66.9% (SVM classifier) to 96.4% (GSA Weighting + SVM) in the Full Chain dataset, from 64.6% (LDA classifier) to 80.2% in the Two Spiral datasets (weighted by BA + LDA). As a result of the study, it was seen that the proposed method successfully fulfills the task of moving the attributes to a linear plane in the datasets, especially in classifiers such as SVM and LDA, which have difficulties in non-linear problems, an accuracy rate of 100% was achieved.

Automated Classification of Whole-Body SPECT Bone Scan Images with VGG-Based Deep Networks

Single Photon Emission Computed Tomography (SPECT) imaging has the potential to acquire information about areas of concerns in a non-invasive manner. Until now, however, deep learning based classification of SPECT images is still not studied yet. To examine the ability of convolutional neural networks on classifying whole-body SPECT bone scan images, in this work, we propose three different two-class classifiers based on the classical Visual Geometry Group (VGG) model. The proposed classifiers are able to automatically identify that whether or not a SPECT image include lesions via classifying this image into categories. Specifically, a pre-processing method is proposed to convert each SPECT file into an image via balancing difference of the detected uptake between SPECT files, normalizing elements of each file into an interval, and splitting an image into batches. Second, different strategies were introduced into the classical VGG 16 model to develop classifiers by minimizing the number of parameters as many as possible. Lastly, a group of clinical whole-body SPECT bone scan files were utilized to evaluate the developed classifiers. Experiment results show that our classifiers are workable for automated classification of SPECT images, obtaining the best values of 0.838, 0.929, 0.966, 0.908 and 0.875 for accuracy, precision, recall, F-1 score and AUC value, respectively.

An Industrial Application Towards Classification and Optimization of Multi-Class Tile Surface Defects Based on Geometric and Wavelet Features

It is possible to detect visual surface defects with software in industrial tile production and increase productivity by automating the quality control process. In this process, low error rate and low cost are important indicators. In order to eliminate this negativity and the effect of the human factor, error detection software has been developed in an artificial intelligence-based industrial artificial vision environment. Spots, scratches, cracks, pore defects, which are the most common surface defects, are classified according to 6 different geometric and wavelet transform attributes. Firstly, an industrial artificial vision environment was created. In this environment, a total of 150 tile images, equal numbers from each class, were obtained on the real-time production line. The resulting images were converted into binary images by preprocessing and filtering. For classification, the support vector machines method, which performs high in two-class classifications, is used with the one versus all approach. In classifications made using RBF kernel function using wavelet features as classification performance, a higher success was achieved in all defect classes than geometric features. Real-time application software for all these processes has been developed with the Python language on Ubuntu operating system.

Deep Learning Based Hybrid Classifier for Analyzing Hepatitis C in Ultrasound Images

Although liver biopsy is the gold standard for identifying diffuse liver disorders, it is an intrusive procedure with a host of negative side effects. Physician subjectivity may affect the ultrasonography diagnosis of diffuse liver disease. As a result, there is still a clear need for an appropriate classification of liver illnesses. In this article, an unique deep classifier made up of deep convolutional neural networks (CNNs) that have already been trained is proposed to categories the liver condition. The variants of ResNet and AlexNet are a few networks that are combined with fully connected networks (FCNs). Transfer learning can be used to extract deep features that can offer adequate categorization data. Then, an FCN can depict images of the disease in its many stages, including tissue, liver hepatitis, and hepatitis. To discriminate between these liver images, three different (normal/cirrhosis, perfectly natural, and cirrhosis/hepatitis) and 3 (normal/cirrhosis/hepatitis) models were trained. A hybrid classifier is presented in order to integrate the graded odds of the classes produced by each individual classifier since two-class classifiers performed better than three-class classifiers. The class with the highest score is then chosen using a majority voting technique. The experimental results demonstrate an high accuracy when liver images were divided into three classes using ResNet50 and a hybrid classifier.

The Optimization of the Light-Source Spectrum Utilizing Neural Networks for Detecting Oral Lesions

Any change in the light-source spectrum modifies the color information of an object. The spectral distribution of the light source can be optimized to enhance specific details of the obtained images; thus, using information-enhanced images is expected to improve the image recognition performance via machine vision. However, no studies have applied light spectrum optimization to reduce the training loss in modern machine vision using deep learning. Therefore, we propose a method for optimizing the light-source spectrum to reduce the training loss using neural networks. A two-class classification of one-vs-rest among the classes, including enamel as a healthy condition and dental lesions, was performed to validate the proposed method. The proposed convolutional neural network-based model, which accepts a 5 × 5 small patch image, was compared with an alternating optimization scheme using a linear-support vector machine that optimizes classification weights and lighting weights separately. Furthermore, it was compared with the proposed neural network-based algorithm, which inputs a pixel and consists of fully connected layers. The results of the five-fold cross-validation revealed that, compared to the previous method, the proposed method improved the F1-score and was superior to the models that were using the immutable standard illuminant D65.

Automatic body condition scoring for dairy cows based on efficient net and convex hull features of point clouds

Body condition score (BCS) is an important indicator to evaluate the nutritional and health status of perinatal dairy cows. The insufficient accuracy of the automatic body condition score is the core bottleneck restricting the popularization and application of this technology. To further improve the accuracy of automatic scoring of cow body condition, a 3D structure feature map based on convex hull distance of point clouds was constructed and used as the input of a two-level model based on EfficientNet network, which realized the improvement of recognition accuracy within the error of 0.25 in the automatic scoring of cow body condition. First, the depth image of the cow’s back was preprocessed to obtain the point cloud from the hook bone to the needle bone region containing the main body condition information. Second, the point cloud was voxelized and convex hulled, and the distance between each peripheral voxel and the nearest convex hull surface was calculated and projected to the X-Y plane to obtain the structural feature map. Then, a two-level classification model based on the EfficientNet network was constructed, and the whale optimization algorithm (WOA) was used to optimize the network scaling coefficient in the model. Finally, 5119 depth images of 77 cows were used to train and test the model, and the dataset ratio was 1:1. The results showed that when the BCS value was in the range of 2.25 to 3.75 and the two-class classification limit was 2.75, the proportions of images with recognition errors less than 0.25 and 0.50 were 91.21% and 97.60%, respectively, and the average recognition speed was 3.446 frames/s. The recognition effect was better than that of the MobileNet-V2, XceptionNet and LeNet-5 network models. This method could realize the noncontact assessment of individual body conditions of dairy cows in large-scale farms and had the characteristics of high accuracy, strong applicability and low cost.

Ensemble feature extraction-based prediction of fetal arrhythmia using cardiotocographic signals

The classification of Cardiotocography (CTG) signal abnormalities is critical in the identification of fetal anomalies. The non-stationary nature of CTG, as well as the dataset asymmetry, make this classification task complicated. Manual CTG analysis interpretation can easily result in a low diagnosis rate, which is mainly caused by a variety of subjective variables. Hence, in this research a new Ensemble Feature Extraction (EFE) model is implemented to analyze the nonlinear cardiotocographic signals. First, the collected signal artifacts are removed with interpolation operation, then down sampled to eliminate the complexity during classification. Followed by, a new EFE module is applied to extract the features, here learning model SE-ResNet-50 (Squeeze and Excitation -ResNet-50) is applied to separate the features from the regular and Time-Frequency data transformed (via CWT transform) signal. Wavelet Packet Transform is exploited to extract T-F signals. Then the enhanced feature set dimension is lessened by using Linear Discriminant Analysis (LDA) model. Finally, the features are fused and fed to the XG-booster for classification, and the patient is advised to seek therapy based on the outcome. The proposed method achieved 96.34% accuracy, 96.7% precision, 97.3% recalls, and 95.64% Quality Index (QI) on two-class classifications. This ensemble feature extraction increases the classification results and proves very advantageous to cardiotocographic signal classification. The results of the experiments show how effective the suggested EFE is.

A hierarchical GAN method with ensemble CNN for accurate nodule detection.

Lung cancer can evolve into one of the deadliest diseases whose early detection is one of the major survival factors. However, early detection is a challenging task due to the unclear structure, shape, and the size of the nodule. Hence, radiologists need automated tools to make accurate decisions. This paper develops a new approach based on generative adversarial network (GAN) architecture for nodule detection to propose a two-step GAN model containing lung segmentation and nodule localization. The first generator comprises a U-net network, while the second utilizes a mask R-CNN. The task of lung segmentation involves a two-class classification of the pixels in each image, categorizing lung pixels in one class and the rest in the other. The classifier becomes imbalanced due to numerous non-lung pixels, decreasing the model performance. This problem is resolved by using the focal loss function for training the generator. Moreover, a new loss function is developed as the nodule localization generator to enhance the diagnosis quality. Discriminator nets are implemented in GANs as an ensemble of convolutional neural networks (ECNNs), using multiple CNNs and connecting their outputs to make a final decision. Several experiments are designed to assess the model on the well-known LUNA dataset. The experiments indicate that the proposed model can reduce the error of the state-of-the-art models on the IoU criterion by about 35 and 16% for lung segmentation and nodule localization, respectively. Unlike recent studies, the proposed method considers two loss functions for generators, further promoting the goal achievements. Moreover, the network of discriminators is regarded as ECNNs, generating rich features for decisions.

Landslide Susceptibility Prediction: Improving the Quality of Landslide Samples by Isolation Forests

Landslide susceptibility prediction (LSP) is the first step to ease landslide disasters with the application of various machine learning methods. A complete landslide inventory, which is essential but difficult to obtain, should include high-quality landslide and non-landslide samples. The insufficient number of landslide samples and the low purity of non-landslide samples limit the performance of the machine learning models. In response, this study aims to explore the effectiveness of isolated forest (IF) to solve the problem of insufficient landslide samples. IF belongs to unsupervised learning, and only a small share of landslide samples in the study area were required for modeling, while the remaining samples were used for testing. Its performance was compared to another advanced integration model, adaptive boosting integrated with decision tree (Ada-DT), which belongs to two-class classifiers (TCC) and needs a sufficient number of samples. Huangpu District, Guangzhou City, Guangdong Province in China, was selected as the study area, and 13 predisposing factors were prepared for the modeling. Results showed that the IF proved its effectiveness with an AUC value of 0.875, although the Ada-DT model performed better (AUC = 0.921). IF outperformed the Ada-DT model in terms of recognizing landslides, and the sensitivity values of IF and the Ada-DT model were 90.00% and 86.67%, respectively, while the Ada-DT model performed better in terms of specificity. Two susceptibility maps obtained by the models were basically consistent with the field investigation, while the areas predicted by IF tended to be conservative as higher risk areas were presented, and the Ada-DT model was likely to be risky. It is suggested to select non-landslide samples from the very low susceptibility areas predicted by the IF model to form a more reliable sample set for Ada-DT modeling. The conclusion confirms the practicality and advancement of the idea of anomaly detection in LSP and improves the application potential of machine learning algorithms for geohazards.

Establishment and evaluation of multiple adulteration detection of camellia oil by mixture design

Multiple adulteration is a common trick to mask adulteration detection methods. In this study, the representative multiple adulterated camellia oils were prepared according to the mixture design. Then, these representative oils were employed to build two-class classification models and validate one-class classification model combined with fatty acid profiles. The cross-validation results indicated that the recursive SVM model possessed higher classification accuracy (97.9%) than PLS-DA. In OCPLS model, the optimal percentage of RO, SO, CO and SUO was 2.8%, 0%, 7.2%, 0% respectively in adulterated camellia oil, which is the most similar to the authentic camellia oils. Further validation showed that five adulterated oils with the optimal percentage could be correctly identified, indicating that the OCPLS model could identify multiple adulterated oils with these four cheaper oils. Moreover, this study serves as a reference for one class classification model evaluation and a solution for multiple adulteration detection of other foods.

Emotion recognition while applying cosmetic cream using deep learning from EEG data; cross-subject analysis.

We report a deep learning-based emotion recognition method using EEG data collected while applying cosmetic creams. Four creams with different textures were randomly applied, and they were divided into two classes, "like (positive)" and "dislike (negative)", according to the preference score given by the subject. We extracted frequency features using well-known frequency bands, i.e., alpha, beta and low and high gamma bands, and then we created a matrix including frequency and spatial information of the EEG data. We developed seven CNN-based models: (1) inception-like CNN with four-band merged input, (2) stacked CNN with four-band merged input, (3) stacked CNN with four-band parallel input, and stacked CNN with single-band input of (4) alpha, (5) beta, (6) low gamma, and (7) high gamma. The models were evaluated by the Leave-One-Subject-Out Cross-Validation method. In like/dislike two-class classification, the average accuracies of all subjects were 73.2%, 75.4%, 73.9%, 68.8%, 68.0%, 70.7%, and 69.7%, respectively. We found that the classification performance is higher when using multi-band features than when using single-band feature. This is the first study to apply a CNN-based deep learning method based on EEG data to evaluate preference for cosmetic creams.

EfficientNet-B0 Model for Face Mask Detection Based on Social Information Retrieval

The world was introduced to the term coronavirus at the end of 2019, following which everyone was thrown into stress and anxiety. The pandemic has been a complete disaster, wreaking devastation and resulting in a significant loss of human life throughout the world. The governments of various countries have issued guidelines and protocols to be followed for stopping the surge in cases (i.e., wearing masks). Amidst all this chaos, the only weapon is technology. So, the detection of face masks is important. The authors utilized a dataset that included images of individuals in society wearing and not wearing masks. They gathered the information required to train a model by using deep networks like EfficientNetB0, MobileNetV2, ResNet50, and InceptionV3. With EfficientNet-B0, they have been able to achieve an accuracy of 99.70% on a two-class classification issue. These methods make face mask detection easier and help in knowledge discovery. These technological breakthroughs may aid in information retrieval as well as help society and guarantee that such a healthcare disaster does not occur again.

BreaST-Net: Multi-Class Classification of Breast Cancer from Histopathological Images Using Ensemble of Swin Transformers

Breast cancer (BC) is one of the deadly forms of cancer, causing mortality worldwide in the female population. The standard imaging procedures for screening BC involve mammography and ultrasonography. However, these imaging procedures cannot differentiate subtypes of benign and malignant cancers. Here, histopathology images could provide better sensitivity toward benign and malignant cancer subtypes. Recently, vision transformers have been gaining attention in medical imaging due to their success in various computer vision tasks. Swin transformer (SwinT) is a variant of vision transformer that works on the concept of non-overlapping shifted windows and is a proven method for various vision detection tasks. Thus, in this study, we investigated the ability of an ensemble of SwinTs in the two-class classification of benign vs. malignant and eight-class classification of four benign and four malignant subtypes, using an openly available BreaKHis dataset containing 7909 histopathology images acquired at different zoom factors of 40×, 100×, 200×, and 400×. The ensemble of SwinTs (including tiny, small, base, and large) demonstrated an average test accuracy of 96.0% for the eight-class and 99.6% for the two-class classification, outperforming all the previous works. Thus, an ensemble of SwinTs could identify BC subtypes using histopathological images and may lead to pathologist relief.

High-Precision Real-Time Forest Fire Video Detection Using One-Class Model

Machine learning-base fire detection methods play a vital role in the current forest fire monitoring systems. In the literature, fire detection is usually viewed as a two-class (or multi-class) classification task. However, this violates the most foundational hypothesis in machine learning, e.g., independent and identical distribution (i.i.d.), especially for the non-fire samples drawn from a complex forest background. Moreover, for omni-directional video -monitoring, the background is also always changing, which leads this violation to a worse situation. In this work, by relaxing the i.i.d. of non-fire samples, we aim to learn a one-class model that just relies on the fire samples. Considering the requirements of high-precision and real-time detection, training samples are directly constructed on the fire pixels without a complex feature transformation. Additionally, we also provide a batch decision-making strategy to speed up fire detection. This work also includes an extensive experimental comparison on the public forest fire videos, obtained by ground- or unmanned aerial vehicle (UAV)-monitoring cameras. Compared with the state-of-the-art methods, the results show the superiority of our proposal in terms of a high-fire detection rate, low-error warning rate, accurate fire location positioning, and real-time detection.

Environmental screening model of driving behavior for an electric bus entering and leaving stops

Framework of the environmental screening model of the driving behavior. Energy consumption determines the environmental benefits and driving range of electric vehicles. The driving behavior significantly affects vehicle energy consumption, but the implicit relationship between them is unclear. Natural driving data of Electric buses (E-Bus) on the BRT line were collected to evaluate the implicit relationship between driving behavior and energy consumption while entering and leaving bus stops. Statistical methods were used to analyze the driving characteristics and their influence on energy consumption. An environmental screening model of the driving behavior was established based on the extreme gradient boosting (XGBoost) algorithm to analyze the impact of the driving parameters on energy consumption and extract their implicit relationship. The results were compared with predictions obtained from random forest (RF) and support vector machine (SVM) models. The results showed that the prediction accuracy of the XGBoost algorithm was higher than that of the RF and SVM models for two-class (95.2%) and three-class (84.9%) classifications. Although the prediction accuracy was lower for the three-class model, its classification was more specific and accurately predicted the energy consumption level.

Differentiation of Motor Speech Disorders through the Seven Deviance Scores from MonPaGe-2.0.s

For the clinical assessment of motor speech disorders (MSDs) in French, the MonPaGe-2.0.s protocol has been shown to be sensitive enough to diagnose mild MSD based on a combination of acoustic and perceptive scores. Here, we go a step further by investigating whether these scores-which capture deviance on intelligibility, articulation, voice, speech rate, maximum phonation time, prosody, diadochokinetic rate-contribute to the differential diagnosis of MSDs. To this aim, we trained decision trees for two-class automatic classification of different pairs of MSD subtypes based on seven deviance scores that are computed in MonPaGe-2.0.s against matched normative data. We included 60 speakers with mild to moderate MSD from six neuropathologies (amyotrophic lateral sclerosis, Wilson, Parkinson and Kennedy disease, spinocerebellar ataxia, post-stroke apraxia of speech). The two-class classifications relied mainly on deviance scores from four speech dimensions and predicted with over 85% accuracy the patient's correct clinical category for ataxic, hypokinetic and flaccid dysarthria; classification of the other groups (apraxia of speech and mixed dysarthria) was slightly lower (79% to 82%). Although not perfect and only tested on small cohorts so far, the classification with deviance scores based on clinically informed features seems promising for MSD assessment and classification.

Classification and Detection of COVID-19 and Other Chest-Related Diseases Using Transfer Learning.

COVID-19 has infected millions of people worldwide over the past few years. The main technique used for COVID-19 detection is reverse transcription, which is expensive, sensitive, and requires medical expertise. X-ray imaging is an alternative and more accessible technique. This study aimed to improve detection accuracy to create a computer-aided diagnostic tool. Combining other artificial intelligence applications techniques with radiological imaging can help detect different diseases. This study proposes a technique for the automatic detection of COVID-19 and other chest-related diseases using digital chest X-ray images of suspected patients by applying transfer learning (TL) algorithms. For this purpose, two balanced datasets, Dataset-1 and Dataset-2, were created by combining four public databases and collecting images from recently published articles. Dataset-1 consisted of 6000 chest X-ray images with 1500 for each class. Dataset-2 consisted of 7200 images with 1200 for each class. To train and test the model, TL with nine pretrained convolutional neural networks (CNNs) was used with augmentation as a preprocessing method. The network was trained to classify using five classifiers: two-class classifier (normal and COVID-19); three-class classifier (normal, COVID-19, and viral pneumonia), four-class classifier (normal, viral pneumonia, COVID-19, and tuberculosis (Tb)), five-class classifier (normal, bacterial pneumonia, COVID-19, Tb, and pneumothorax), and six-class classifier (normal, bacterial pneumonia, COVID-19, viral pneumonia, Tb, and pneumothorax). For two, three, four, five, and six classes, our model achieved a maximum accuracy of 99.83, 98.11, 97.00, 94.66, and 87.29%, respectively.

MER-ESTIMATOR OF MULTIDIMENSIONAL BAYESIAN THRESHOLD IN TWO-CLASS CLASSIFICATION PROBLEM

MER-ESTIMATOR OF MULTIDIMENSIONAL BAYESIAN THRESHOLD IN TWO-CLASS CLASSIFICATION PROBLEM

Detection of cervical precancerous cells from Pap-smear images using ensemble classification

Background. Cervical cancer begins in superficial cells and over time can invade deeper tissues and surrounding tissues. This paper presents a creative idea of using an ensemble classification algorithm that improves the predictive performance of an artificial intelligence system based on cervical cancer screening. This study aimed to classify Pap-smear images by different machine learning methods to achieve high accuracy detection. Methods. This study was performed on 917 Pap-smear images from the Herlev public database. In the feature extraction stage, 20 geometric features and 76 texture features were extracted. After that, using ensemble classification method, the images were classified into two categories (i.e., normal and abnormal) and then into seven categories (i.e., superficial epithelial, intermediate epithelial, columnar epithelial, mild dysplasia, moderate dysplasia, severe dysplasia and carcinoma) and the accuracy of the proposed method was evaluated. Results. The algorithm in the ensemble classification was able to achieve accuracy of 99.9% with a processing time of 0.028 second in the two-class classification and accuracy of 76.5% with a processing time of 0.033 second in the seven-class classification. Conclusion. Based on the results, the designed algorithm can be used as a computer aided diagnostic tool to increase the accuracy and speed of predicting the risk of cervical cancer. Practical Implications. Cervical cancer is one of the most common cancers among women. Early diagnosis of the disease can save various costs and prevent the patients’ frequent visits to medical centers. This research proposed an artificial intelligence method for automatic classification of cervical cells and improving the accuracy of diagnosis.

A Method for the Study of Cerebellar Cognitive Function-Re-Cognition and Validation of Error-Related Potentials.

The cerebellar region has four times as many brain cells as the brain, but whether the cerebellum functions in cognition, and how it does so, remain unexplored. In order to verify whether the cerebellum is involved in cognition, we chose to investigate whether the cerebellum is involved in the process of error judgment. We designed an experiment in which we could activate the subject’s error-related potentials (ErrP). We recruited 26 subjects and asked them to wear EEG caps with cerebellar regions designed by us to participate in the experiment so that we could record their EEG activity throughout the experiment. We successfully mitigated the majority of noise interference after a series of pre-processing of the data collected from each subject. Our analysis of the preprocessed data revealed that our experiment successfully activated ErrP, and that the EEG signals, including the cerebellum, were significantly different when subjects made errors compared to when they made correct judgments. We designed a feature extraction method that requires selecting channels with large differences under different classifications, firstly by extracting the time-frequency features of these channels, and then screening these features with sequence backward feature (SBS) selection. We use the extracted features as the input and different event types in EEG data as the labels for multiple classifiers to classify the data in the executive and feedback segments, where the average accuracy for two-class classification of executive segments can reach 80.5%. The major contribution of our study is the discovery of the presence of ErrP in cerebellar regions and the extraction of an effective feature extraction method for EEG data.