One-class Classification Approach Research Articles

Comprehensive Review of One-Class Classification Approaches for Anomaly Detection

Comprehensive Review of One-Class Classification Approaches for Anomaly Detection

Metabolomic analysis to predict the onset and severity of necrotizing enterocolitis

BackgroundNecrotizing enterocolitis (NEC) is the most devastating gastrointestinal (GI) emergency in preterm neonates. Untargeted metabolomics may allow the identification of biomarkers involved in NEC pathophysiology.MethodsWe conducted a prospective study including preterm infants born at < 34 gestational weeks (GWs) whose urine was longitudinally collected at birth (< 48 h, T0) and at 14 (T1) and 28 days (T2). Neonates were followed for their development of NEC, spontaneous intestinal perforation (SIP), or other GI conditions and compared to those of matched healthy controls. Urine samples were investigated by untargeted metabolomic analysis based on mass-spectrometry.ResultsThirty-five patients with NEC, 5 patients with SIP, 14 patients with other GI diseases and 113 controls were enrolled and selected for metabolomic analysis on the basis of their clinical characteristics and available samples. Considering urine samples at T0, the one-class classification approach was able to correctly classify 16/20 subjects (80%) who developed NEC, 3/3 (100%) who developed SIP and 5/7 subjects (71.4%) with other GI pathologies as not belonging to the control group. Neonates with surgical NEC had higher N-acetylaspartic acid, butyrylcarnitine and propionylcarnitine levels than did those with medical NEC. Considering the time evolution of the urinary metabolome, the NEC and control groups showed differences independently of the time point.ConclusionsThe urinary metabolome is closely associated with the underlying GI disease from birth. Urinary metabolic features characterize NEC patients from healthy controls until 28 days of life. The early urinary metabolome has the potential to predict surgical NEC. Future studies are needed to validate our results.

Robust one-class classification using deep kernel spectral regression

The existing one-class classification (OCC) methods typically presume the existence of a pure target training set and generally face difficulties when the training set is contaminated with non-target objects. This work addresses this aspect of the OCC problem and formulates an effective method that leverages the advantages of kernel-based methods to achieve robustness against training label noise while enabling direct deep learning of features from the data to optimise a Fisher-based loss function in the Hilbert space. As such, the proposed OCC approach can be trained in an end-to-end fashion while, by virtue of a Tikhonov regularisation in the Hilbert space, it provides high robustness against the training set contamination.Extensive experiments conducted on multiple datasets in different application scenarios demonstrate that the proposed methodology is robust and performs better than the state-of-the-art algorithms for OCC when the training set is corrupted by contamination.

Privileged multi-view one-class support vector machine

One-class support vector machine (OCSVM) is a typical one-class classification approach, which learns the classifier by using only the target samples. At present, most OCSVM works hypothesize that the samples have only one view, while multi-view OCSVM has not been taken into account. In this paper, a novel multi-view one-class support vector machine method with privileged information learning (MOCPIL) is put forward. MOCPIL embodies both the consensus principle and complementarity principle in multi-view learning. Privileged information is additional data that is available only in the training process, but not in the testing process. By introducing the idea of privileged information learning, MOCPIL implements the complementarity principle by treating one view as the training data and the other view as the privileged data. Moreover, MOCPIL implements the consensus principle by requiring that different views of the same object should give similar predicting outputs. The learning problem of MOCPIL is a quadratic programming (QP) problem, which is able to be solved by off-the-shelf QP solvers. To the best of our knowledge, this is the first study to tackle the multi-view learning problem based on OCSVM. The performance of MOCPIL is evaluated through extensive experiments. The experimental results have shown that MOCPIL explicitly outperforms the existing multi-view one-class classification methods.

A one-class classification approach for authentication of specialty coffees by inductively coupled plasma mass spectroscopy (ICP-MS)

Due to the lucrative nature of specialty coffees, there have been instances of adulteration where low-cost materials are mixed in to increase the overall volume, resulting in illegal profit. A widely used and recommended approach to detect possible adulteration is the application of one-class classifiers (OCC), which only require information about the target class to build the models. Thus, this work aimed to identify adulterations in specialty coffees with low-quality coffee using multielement analysis determined by ICP-MS and to evaluate the performance of one-class classifiers (dd-SIMCA, OCRF, and OCPLS). Therefore, authentic specialty coffee samples were adulterated with low-quality coffee in 25 % to 75 % (w/w) proportions. Samples were subjected to acid decomposition for analysis by ICP-MS. OCPLS method presented the best performance to detect adulterations with low-quality coffee in specialty coffees, showing higher specificity (SPE = 100 %) and reliability rate (RLR = 94.3 %).

Imbalanced Anomaly Detection Using Augmented Deeper FCDDs for Wooden Sleeper Deterioration Prognostics

Maintaining high standards for user safety during daily railway operations is crucial for railway managers. To aid in this endeavor, top or side-view cameras and GPS positioning systems have facilitated progress toward automating periodic inspections of defective features and assessing the deteriorating status of railway components. However, collecting data on deteriorated status can be time-consuming and requires repeated data acquisition because of the extreme temporal occurrence imbalance. In supervised learning, thousands of paired data sets containing defective raw images and annotated labels are required. Concretely, the one-class classification approach offers the advantage of requiring quite a few anomalous images to optimize parameters for training large normal images. The deeper fully-convolutional data descriptions (FCDDs) were applicable to several damage data sets of concrete/steel components in structures, and fallen tree, and wooden building collapse in disasters. However, it is not yet known to feasible to railway components. In this study, we devised a prognostic discriminator pipeline to automate one class classification using the augmented deeper FCDDs for defective railway components. We also performed sensitivity analysis of the mixture and erasing augmentations, and the deeper backbone rather than the shallow baseline of convolutional neural network (CNN) with 27 layers. Furthermore, we visualized defective railway features by using transposed Gaussian upsampling. We demonstrated our application to railway inspection using a video acquisition dataset that contains wooden sleeper deterioration. Finally, we examined the usability of our approach for prognostic monitoring and fu ture work on railway component inspection.

Global Plus Local Jointly Regularized Support Vector Data Description for Novelty Detection.

In many practice application, the cost for acquiring abnormal data is quite expensive, thus the one-class classification (OCC) problem attracts great attention. As one of the solutions, support vector data description (SVDD) gains a continuous focus in outlier detection since it is based on the data description. For the sphere obtained by SVDD, both the center and the volume (or radius) strongly depend on the support vectors, while the support vectors are sensitive to the tradeoff parameter C. Hence, how to select this parameter is a rather challenging problem. In order to address this problem, we define several distance metrics relative to the image region in Gaussian kernel space. With the distance metrics, two probability densities relative to the global region and the local region are designed, respectively. Then, the information quantity and the information entropy are developed for regularizing the tradeoff parameter. This novel SVDD is called global plus local jointly regularized support vector data description (GL-SVDD), in which both the global region information and the local image region information jointly penalize the images as possible outliers. Finally, we use the UCI dataset and the hyperspectral data of cherry fruit to evaluate the performance of several OCC approaches. Experimental results show that GL-SVDD is encouraging.

Learning Deep Hierarchical Features with Spatial Regularization for One-Class Facial Expression Recognition

Existing methods on facial expression recognition (FER) are mainly trained in the setting when multi-class data is available. However, to detect the alien expressions that are absent during training, this type of methods cannot work. To address this problem, we develop a Hierarchical Spatial One Class Facial Expression Recognition Network (HS-OCFER) which can construct the decision boundary of a given expression class (called normal class) by training on only one-class data. Specifically, HS-OCFER consists of three novel components. First, hierarchical bottleneck modules are proposed to enrich the representation power of the model and extract detailed feature hierarchy from different levels. Second, multi-scale spatial regularization with facial geometric information is employed to guide the feature extraction towards emotional facial representations and prevent the model from overfitting extraneous disturbing factors. Third, compact intra-class variation is adopted to separate the normal class from alien classes in the decision space. Extensive evaluations on 4 typical FER datasets from both laboratory and wild scenarios show that our method consistently outperforms state-of-the-art One-Class Classification (OCC) approaches.

OCSTN: One-class time-series classification approach using a signal transformation network into a goal signal

One-class classification (OCC) is a classification task where the training data have only one class. The goal is to classify input data into one seen class or other unseen classes. This paper proposes an OCC approach using a signal transformation network (OCSTN), which aims to process univariate time-series data. The main contribution is developing a signal transformation network (STN) that aims to transform input signals into one signal, namely the goal signal. Moreover, the model error of the STN is a distance metric between the goal signal and the model output. The STN model learns from one-class signals. Therefore, model error for one class is small relative to other classes. Accordingly, OCSTN could discriminate between seen and unseen classes using the model errors. The proposed OCSTN is evaluated using two ballistocardiography (BCG) datasets. The OCSTN achieves fair results in both AUC scores and processing speed. OCSTN has a weak point in training diverse signals. In addition, the entropy and smoothness of the goal signal are highly related to the AUC score.

Deep learning based detection and localization of road accidents from traffic surveillance videos

Real-world traffic surveillance videos need continuous supervision to monitor and take appropriate actions in case of fatal accidents. However, continuously monitoring them with human supervision is error prone and tedious. Therefore, a deep learning approach for automatic detection and localization of road accidents has been proposed by formulating the problem as anomaly detection. The method follows one-class classification approach and applies spatio-temporal autoencoder and sequence-to-sequence long short-term memory autoencoder for modeling spatial and temporal representations in the video. The model is executed on a real-world video traffic surveillance datasets and significant results have been achieved both qualitatively and quantitatively.

OCFSP: self-supervised one-class classification approach using feature-slide prediction subtask for feature data

One-class classification (OCC) is a machine learning problem where training data has only one class. Recently, self-supervised OCC algorithms have been increasing attention. These algorithms train the model for pretext tasks and use the model error for OCC. However, these tasks are specialized for images, and applying them to feature data is not practical or appropriate for such a purpose. The motivation of this study is to apply self-supervised OCC to feature data. For this purpose, this paper proposes an OCC approach using feature-slide prediction (FSP) subtask for feature data (OCFSP). The main originality is the FSP subtask, which is the first classification subtask for feature data. In particular, the proposed method creates a self-labeled dataset by generating additional feature vectors with the feature slide of original vectors and self-annotating these vectors as the number of the slides. Such a dataset is applied to train a multi-class classifier to predict the number of feature slides. Since this classification model learns data from only one class, the FSP accuracy for a seen class is higher relative to unseen classes. Accordingly, OCC could be made using the accuracy of FSP. The proposed methods are experimented with using the imbalanced-learn, covtype, and kddcup datasets. OCFSP shows fair accuracy where few training data is given. In addition, classification subtask for feature data shows a relatively fast testing speed, unlike image data. Therefore, the bottleneck of the self-supervised approach is considered the memory size, which is the main difference between image and feature data. Source code is uploaded at https://github.com/ToshiHayashi/OCFSP

Assessment of meat authenticity using portable Fourier transform infrared spectroscopy combined with multivariate classification techniques

In the present contribution, the feasibility of portable Fourier transform infrared spectroscopy (FTIR) combined with multivariate classification techniques is assessed for classification of minced beef, lamb, chicken and pork samples. In this regard, both attenuated total reflectance-FTIR (ATR-FTIR) and diffuse reflectance-FTIR (DR-FTIR) methods are evaluated. First, principal component analysis (PCA) was used for exploring FT-IR spectra of four meat species to find similarities and dissimilarities among samples. Additionally, one-class classification (OCC) was utilized as a new approach for halal meat species certification. For OCC, two scenarios were defined: (i) 100% correct classification for pork, and (ii) a most favorable overall classification rate for all species investigated simultaneously. With the OCC approach, both ATR and DR methods were found to produce high false-positive scores in scenario (i), whilst the DR method scored the best in scenario (ii) with an overall score of 89% correct classification. In the next step, partial least squares-discriminant analysis (PLS-DA) and support vector machine (SVM) with radial basis function (RBF) as kernel function were evaluated for meat speciation. On this matter, SVM showed better classification performance in terms of total accuracy for both ATR-FTIR (98%) and DR-FTIR (100%) datasets over PLS-DA (90% and 98%, respectively).The promising results of both portable ATR-FTIR and DR-FTIR combined with OCC approach and discriminant analysis indicated for the first time their use as successful non-destructive, cost-effective and rapid routine screening methods for on-site analysis of meat speciation and halal meat species certification which could be useful for quality control officers to manage and control meat authenticity at various stages of the supply chain.

Identifying large-area mangrove distribution based on remote sensing: A binary classification approach considering subclasses of non-mangroves

Mangroves have tremendous ecological value but are vulnerable to anthropogenic factors and sea level rise. Classification based on remote sensing is the first step for monitoring mangrove distribution in large areas but suffers from misclassifications due to the complexity of mangroves, such as having abundant species within the mangrove class, a large latitudinal span from tropical to subtropical, and mixing with nearby forests at the fringe. Binary classification approaches (i.e., mangroves and non-mangroves) have the potential to adjust the decision surface by adding subclasses similar to the mangrove class into the non-mangrove class in comparison with one-class classification approaches (i.e., separating mangroves from the background), and it demands fewer samples than multi-class classification (i.e., a mangrove class and multiple non-mangrove classes). However, the subclasses of binary classifications have rarely been considered to reduce the sample size with comparable performance to multi-class classification approaches. This paper first tested whether a conversion from either of two existing multi-class classifications to a binary classification could achieve comparable performance to corresponding multi-class classification for identifying mangroves in China in 2018 using the Google Earth Engine platform. The results showed that the conversion from multi-class classifications to binary classification achieved an overall accuracy value of 74.1% and 83.8%, respectively, the second of which was higher than that of the corresponding multi-class classification result with a reduction of sample size by a rate of 72.9%. Such an improved performance is due to the adjustment of subclasses during conversion when evaluated by the same validation dataset. Thus, this study provided a binary classification approach to identifying large-area mangrove distribution with a reduced sample size and comparable performance. This paper further discussed the impact of the feature selection procedure and that of subclass settings to reveal the margins of the proposed approach, the distribution and causes of false positives in binary classification results, and adaptations suitable for further generalizations to case areas outside of China. This study fills a knowledge gap and serves for accurate and consistent large-area mangrove identification based on remote sensing.

Multiscale generative model using regularized skip-connections and perceptual loss for anomaly detection in toxicologic histopathology

BackgroundAutomated anomaly detection is an important tool that has been developed for many real-world applications, including security systems, industrial inspection, and medical diagnostics. Despite extensive use of machine learning for anomaly detection in these varied contexts, it is challenging to generalize and apply these methods to complex tasks such as toxicologic histopathology (TOXPATH) assessment (i.e.,finding abnormalities in organ tissues). In this work, we introduce an anomaly detection method using deep learning that greatly improves model generalizability to TOXPATH data. MethodsWe evaluated a one-class classification approach that leverages novel regularization and perceptual techniques within generative adversarial network (GAN) and autoencoder architectures to accurately detect anomalous histopathological findings of varying degrees of complexity. We also utilized multiscale contextual data and conducted a thorough ablation study to demonstrate the efficacy of our method. We trained our models on data from normal whole slide images (WSIs) of rat liver sections and validated on WSIs from three anomalous classes. Anomaly scores are collated into heatmaps to localize anomalies within WSIs and provide human-interpretable results. ResultsOur method achieves 0.953 area under the receiver operating characteristic on a real-worldTOXPATH dataset. The model also shows good performance at detecting a wide variety of anomalies demonstrating our method’s ability to generalize to TOXPATH data. ConclusionAnomalies in both TOXPATH histological and non-histological datasets were accurately identified with our method, which was only trained with normal data.

The Feasibility of Two Handheld Spectrometers for Meat Speciation Combined with Chemometric Methods and Its Application for Halal Certification.

Handheld visible-near-infrared (Vis-NIR) and near-infrared (NIR) spectroscopy can be cost-effective, rapid, non-destructive and transportable techniques for identifying meat species and may be valuable for enforcement authorities, retail and consumers. In this study, a handheld Vis-NIR (400–1000 nm) and a handheld NIR (900–1700 nm) spectrometer were applied to discriminate halal meat species from pork (halal certification), as well as speciation of intact and ground lamb, beef, chicken and pork (160 meat samples). Several types of class modeling multivariate approaches were applied. The presented one-class classification (OCC) approach, especially with the Vis-NIR sensor (95–100% correct classification rate), was found to be suitable for the application of halal from non-halal meat-species discrimination. In a discriminant approach, using the Vis-NIR data and support vector machine (SVM) classification, the four meat species tested could be classified with accuracies of 93.4% and 94.7% for ground and intact meat, respectively, while with partial least-squares discriminant analysis (PLS-DA), classification accuracies were 87.4% (ground) and 88.6% (intact). Using the NIR sensor, total accuracies of the SVM models were 88.2% and 81.5% for ground and intact meats, respectively, and PLS-DA classification accuracies were 88.3% (ground) and 80% (intact). We conclude that the Vis-NIR sensor was most successful in the halal certification (OCC approaches) and speciation (discriminant approaches) for both intact and ground meat using SVM.

Class-Modeling of Septic Shock With Hyperdimensional Computing.

Sepsis arises when a patient's immune system has an extreme reaction to an infection. This is followed by septic shock if damage to organ tissue is so extensive that it causes a total systemic failure. Early detection of septic shock among septic patients could save critical time for preparation and prevention treatment. Due to the high variance in symptoms and patient state before shock, it is challenging to create a protocol that would be effective across patients. However, since septic shock is an acute change in patient state, modeling patient stability could be more effective in detecting a condition that departs from it. In this paper we present a one-class classification approach to septic shock using hyperdimensional computing. We built various models that consider different contexts and can be adapted according to a target priority. Among septic patients, the models can detect septic shock accurately with 90% sensitivity and overall accuracy of 60% of the cases up to three hours before the onset of septic shock, with the ability to adjust predictions according to incoming data. Additionally, the models can be easily adapted to prioritize sensitivity (increase true positives) or specificity (decrease false positives).

An assessment of Planet satellite imagery for county-wide mapping of rice planting areas in Jiangsu Province, China with one-class classification approaches

ABSTRACT There is a growing demand of the spatial distribution of major crops at the field level in recent years for the precision management of farmlands in developing countries. However, most previous studies used low or medium resolution images (e.g. Moderate Resolution Imaging Spectroradiometer (MODIS) and Landsat) to map the planting areas of rice in various regions of China. As a new source of satellite imagery, the 3 m Planet satellite imagery with daily revisit frequency has rarely been used for mapping crop types. In particular, it is still unclear how well the Planet imagery would perform in mapping the rice planting areas in the Lower Reaches of the Yangtze River Plain of China where image acquisition is strongly limited by rainy and cloudy weather conditions. To overcome the spectral confusion between rice and non-rice vegetation types, this study proposed one-class classification approaches with Planet imagery from the early and late seasons of rice crops over three counties (Guanyun, Rugao and Wujiang) across Jiangsu Province in eastern China. Apart from the traditional use of mono-temporal imagery from the late season (Strategy I), we developed two strategies to make use of the bi-temporal imagery from the early and late seasons, with one being direct classification using stacked images (Strategy II) and the other being stepwise classification (Strategy III). All classifications were performed with the popular machine learning classifier one-class support vector machine (OCSVM) and evaluated from the aspects of classification accuracy and area estimation accuracy. The results showed Strategy I could achieve an overall classification accuracy of 83.02% and kappa coefficient (ƙ) of 0.65. The overall accuracy (OA) can be significantly improved by Strategy II (OA = 84.44% and ƙ = 0.69) and further by Strategy III (OA = 90.42% and ƙ = 0.81). In terms of area estimation, the estimates from Strategy III were consistent with the government reported statistical data with a relative error (RE) within 10% at the county level. At the town level, the correspondence between Planet-derived rice planting areas and the Landsat-8 derived reference data was significantly improved with Strategy III as compared to the other two for Guanyun and Wujiang. The stepwise classification approach (Strategy III) developed with Planet imagery has great values in monitoring rice paddies over the regions with fragmentized distribution of rice fields.

Less complexity one-class classification approach using construction error of convolutional image transformation network

One-class classification is a machine learning problem, where training data has only one class. The objective is to determine if the input data is seen class or unseen class. Traditional deep learning algorithms are not suitable for this task since the algorithm can predict only class in training data. In this paper, the one-class classification algorithm using construction error of image transformation network (OCITN) is proposed. In particular, image transformation network (ITN) is introduced as a subtask, which transforms input image into one image, namely goal image. Moreover, the error of ITN, namely construction error (CE), is computed as a distance metric between the goal image and model output. ITN model is trained using only one-class images and is applied for testing images. Since the model is trained with only one-class images, the CE for one-class is small relative to other classes. Thus, one-class classification is made determining CE is large or small. The proposed method is experimented with using MNIST, Fashion MNIST, CIFAR10, CIFAR100, and Cat-vs-Dog datasets. OCITN shows good results where the goal image has high entropy. Additionally, the extension of OCITN, namely OCITNE, is implemented. This method shows the state of the art performance in MNIST (98.0), Fashion MNIST(95.6), and acceptable performance in CIFAR10(78.4). Furthermore, these methods provide high-speed processing, OCITN process 5291 images, and OCITNE 1261 images per second, 137 times and 33 times faster than state of the art. The source code used in this paper can be downloaded from: https://github.com/ToshiHayashi/OCITN.

Depression and anorexia detection in social media as a one-class classification problem

Taking advantage of the increasing amount of user-generated content in social media, some computational methods have already been proposed for detecting people suffering from depression and anorexia. Such complex tasks have been tackled as a binary classification problem using, in most cases, automatically generated training data. Despite its promising results, this approach has some important drawbacks, namely: it suffers from a severely skewed class distribution, the negative class is very diverse since it attempts to model all kinds of healthy users, and, above all, there is not a complete certainty about annotations, especially for the negative cases (i.e., healthy users). Motivated by these issues, in this paper, we propose to face the detection of these disorders following a one-class classification (OCC) approach. Particularly, we introduce two new instance-based OCC methods especially suited to manage the high diversity of content from social media documents. Taking up ideas from the gravitational attraction force, these methods evaluate the relation of documents by their strengths, considering their distances as well as their masses (relevance) with respect to the target task. Experiments were conducted on depression and anorexia benchmark datasets. The obtained results are encouraging; the overall performance was better than the results from other standard OCC methods, and competitive with regard to state-of-the-art results from binary classification approaches.

A One-class Classification Approach Based on SVDD for Imbalanced and Overlapping Classes

Imbalanced data classification is a challenge especially when there is overlapping between two classes. The overlap makes it almost impossible to create differences in the two classes and isolate them. In the real world, many of the datasets are imbalanced and overlapped. This paper identifies the overlapping regions optimally by comparing the results of a single-class support vector data description (SVDD) algorithm performed on each class. Then, the method uses the nearest-neighbour algorithm to classify the data in the overlapping region. The result of the evaluation on the datasets with a high imbalanced rate shows better performance than other approaches.