Severe Class Research Articles

Intrinsic Consistency Preservation With Adaptively Reliable Samples for Source-Free Domain Adaptation.

Unsupervised domain adaptation (UDA) aims to alleviate the domain shift by transferring knowledge learned from a labeled source dataset to an unlabeled target domain. Although UDA has seen promising progress recently, it requires access to data from both domains, making it problematic in source data-absent scenarios. In this article, we investigate a practical task source-free domain adaptation (SFDA) that alleviates the limitations of the widely studied UDA in simultaneously acquiring source and target data. In addition, we further study the imbalanced SFDA (ISFDA) problem, which addresses the intra-domain class imbalance and inter-domain label shift in SFDA. We observe two key issues in SFDA that: 1) target data form clusters in the representation space regardless of whether the target data points are aligned with the source classifier and 2) target samples with higher classification confidence are more reliable and have less variation in their classification confidence during adaptation. Motivated by these observations, we propose a unified method, named intrinsic consistency preservation with adaptively reliable samples (ICPR), to jointly cope with SFDA and ISFDA. Specifically, ICPR first encourages the intrinsic consistency in the predictions of neighbors for unlabeled samples with weak augmentation (standard flip-and-shift), regardless of their reliability. ICPR then generates strongly augmented views specifically for adaptively selected reliable samples and is trained to fix the intrinsic consistency between weakly and strongly augmented views of the same image concerning predictions of neighbors and their own. Additionally, we propose to use a prototype-like classifier to avoid the classification confusion caused by severe intra-domain class imbalance and inter-domain label shift. We demonstrate the effectiveness and general applicability of ICPR on six benchmarks of both SFDA and ISFDA tasks. The reproducible code of our proposed ICPR method is available at https://github.com/CFM-MSG/Code_ICPR.

One-Shot Weakly-Supervised Segmentation in 3D Medical Images.

Deep neural networks typically require accurate and a large number of annotations to achieve outstanding performance in medical image segmentation. One-shot and weakly-supervised learning are promising research directions that reduce labeling effort by learning a new class from only one annotated image and using coarse labels instead, respectively. In this work, we present an innovative framework for 3D medical image segmentation with one-shot and weakly-supervised settings. Firstly a propagation-reconstruction network is proposed to propagate scribbles from one annotated volume to unlabeled 3D images based on the assumption that anatomical patterns in different human bodies are similar. Then a multi-level similarity denoising module is designed to refine the scribbles based on embeddings from anatomical- to pixel-level. After expanding the scribbles to pseudo masks, we observe the miss-classified voxels mainly occur at the border region and propose to extract self-support prototypes for the specific refinement. Based on these weakly-supervised segmentation results, we further train a segmentation model for the new class with the noisy label training strategy. Experiments on three CT and one MRI datasets show the proposed method obtains significant improvement over the state-of-the-art methods and performs robustly even under severe class imbalance and low contrast. Code is publicly available at https://github.com/LWHYC/OneShot_WeaklySeg.

The treatment of severe class II division 2 patients with deep overbites and gummy smiles using lingual appliances: a case series

Abstract The treatment of Class II division 2 patients with palatally inclined upper incisors using lingual appliances is generally challenging due to difficulties in achieving upper incisor torque control. The present case series aims to describe a simple method combining lingual appliances and mini-screws for the non-surgical management of two severe Class II division 2 patients presenting with deep overbites and gummy smiles. The treatment plans involved initial bite opening and upper incisor torque recovery, followed by distalisation of the entire upper arch using mini-screw anchorage. In one case, additional anterior labial mini-screws were necessary to assist incisor intrusion and complete upper arch levelling. Favourable treatment results suggest that lingual appliances, when combined with mini-screws and pre-torqued arch wires, may offer an effective approach for the non-surgical management of severe Class II division 2 patients with deep overbites and gummy smiles.

A Successful Ventilatory Approach in a Severe Class 3 Obesity Patient Undergoing Robotic Gynecological Surgery

A Successful Ventilatory Approach in a Severe Class 3 Obesity Patient Undergoing Robotic Gynecological Surgery

A novel wood preservative with plant extracts in a cypermethrin mixture protects envelope-treated tropical kempas hardwood against Coptotermes termites when exposed to an above-ground indoor situation after evaporative ageing

In keeping with sustainability initiatives in the commercial wood protection sector to develop cost-effective, environmentally acceptable wood treatment alternatives to traditional biocides, this paper reports on an above-ground test for protection against subterranean termite damage to blocks of termite-susceptible heartwood of tropical (Malaysian) kempas hardwood (Koompassia malaccensis). The blocks were envelope-treated (brush-on) with a novel wood preservative based on plant extracts in a cypermethrin mixture (Biocide 1: 0.16% cypermethrin, 0.08% tebuconazole, 2% vegetal extracts). A commercial permethrin-based LOSP (Biocide 2: 0.2% permethrin, 1.8% tributyltin naphthenate, 0.1% dichlorofluanid) applied in 3-minute dips served as the control treatment. The air-dried kempas blocks were H2-weathered under a rigorous laboratory volatilization regime (representing long-term evaporative ageing of treated wood for above-ground indoor use in prolonged H2-hazard class situations). The blocks were then exposed for 6 months above ground, inside a containerised test design targeting subterranean Coptotermes curvignathus termites, at a humid forest site (representing a severe H2-hazard class situation with treated wood isolated from wetting, soil contact, and weather). The results showed irrefutably that untreated kempas heartwood was severely attacked by C. curvignathus (mean mass loss: 70.4% and 20,416 mg) with a low mean visual termite rating (2.4). However, the treated kempas wood was well protected at very low surface retention of Biocide 1 (mean negligible mass loss: 0.66% and 207 mg) and very low surface retention of the control Biocide 2 (mean negligible mass loss: 1.01% and 306 mg), with both treatments similarly yielding the highest mean visual termite rating for all replicate kempas specimens (rating 10). The performances of the two biocides were similar but varied significantly (P < 0.05) from the attacks against their untreated counterparts. Biocide 1 therefore has considerable anti-termite potential and could replace conventional LOSP Biocide 2 treatments to protect wood exposed to a long-term H2-hazard class environment.

Trans-Balance: Reducing demographic disparity for prediction models in the presence of class imbalance

Introduction:Risk prediction, including early disease detection, prevention, and intervention, is essential to precision medicine. However, systematic bias in risk estimation caused by heterogeneity across different demographic groups can lead to inappropriate or misinformed treatment decisions. In addition, low incidence (class-imbalance) outcomes negatively impact the classification performance of many standard learning algorithms which further exacerbates the racial disparity issues. Therefore, it is crucial to improve the performance of statistical and machine learning models in underrepresented populations in the presence of heavy class imbalance. Method:To address demographic disparity in the presence of class imbalance, we develop a novel framework, Trans-Balance, by leveraging recent advances in imbalance learning, transfer learning, and federated learning. We consider a practical setting where data from multiple sites are stored locally under privacy constraints. Results:We show that the proposed Trans-Balance framework improves upon existing approaches by explicitly accounting for heterogeneity across demographic subgroups and cohorts. We demonstrate the feasibility and validity of our methods through numerical experiments and a real application to a multi-cohort study with data from participants of four large, NIH-funded cohorts for stroke risk prediction. Conclusion:Our findings indicate that the Trans-Balance approach significantly improves predictive performance, especially in scenarios marked by severe class imbalance and demographic disparity. Given its versatility and effectiveness, Trans-Balance offers a valuable contribution to enhancing risk prediction in biomedical research and related fields.

Model for classification of heart failure severity in patients with hypertrophic cardiomyopathy using a deep neural network algorithm with a 12-lead electrocardiogram

ObjectivesIn hypertrophic cardiomyopathy (HCM), specific ECG abnormalities are observed. Therefore, ECG is a valuable screening tool. Although several studies have reported on estimating the risk of developing fatal arrhythmias from...

Multi-Scale Feature Pyramid for Detection of Red Lesions in Fundus Images

Diabetic retinopathy (DR) is increasing rapidly around the world, but there is a shortage of experienced ophthalmologists. Therefore, computer-based diagnosis of the fundus images is essential to screening of referable DR. Automated detection of red lesions is very important for screening of DR. This paper deals with a novel method for automatic detection of red lesion. The main contribution is developing a deep learning based detection framework to handle severe class imbalance and imbalance in sizes of red lesions. The multi-scale features are extracted using the feature pyramid network. A pyramid of features is generated with strong semantics. The proposed network is end-to-end trainable in image level with several scales and works for a wide range of red lesions with acceptable performance. Sensitivity of the proposed method is 0.76 with six false-positive per image on test set of publicly available DIARECTDB1 database and outperforms state-of-the-art approaches. A potential benefit with deep learning based detection framework could be used in screening programs of referable DR.

BIAS: Bridging Inactive and Active Samples for active source free domain adaptation

As a practical domain adaptation scenario for data protection, source free domain adaptation has suffered from a performance bottleneck. Active source free domain adaptation (ASFDA) could alleviate this problem by exploring a few actively labeled target samples. Due to active selection bias in ASFDA, the active samples are hard samples as they are challenging to recognize by the source model, and the inactive samples are composed of well-learned not-hard samples and unexplored mild-hard samples. Thus the distributions of inactive and active samples are usually different. We provide a rigorous theoretical analysis, revealing that reducing their distribution difference is critical for model generalization. However, the previous DA alignment methods cannot reduce such distribution discrepancy due to severe class imbalance and quantity imbalance between the inactive and active samples. In this work, we propose Bridging Inactive and Active Samples (BIAS) to implicitly align their distributions by utilizing their distribution characteristics. By constructing and learning the intermediate samples between the reliable not-hard and hard samples obtained by our novel reciprocal active selection, BIAS can effectively align the distributions of inactive and active samples and learn unexplored mild-hard samples thoroughly. Extensive experiments verify that BIAS can be integrated into existing ASFDA methods and improve their performance significantly, with over 4% accuracy improvements in some transfer tasks.

Imbalanced domain generalization via Semantic-Discriminative augmentation for intelligent fault diagnosis

Domain generalization-based fault diagnosis (DGFD) has garnered significant attention due to its ability to generalize prior diagnostic knowledge to unseen working conditions or machines. However, existing DGFD methods are generally implemented under the premise of class balance, which may not accurately reflect real-world diagnosis scenarios since fault data collected in practical engineering often exhibits severe class imbalance. To address this challenge, this paper proposes a semantic-discriminative augmentation-driven network for imbalanced domain generalization fault diagnosis. A semantic regularization-based mixup strategy is devised to synthesize sufficient reliable samples to compensate for minority classes. Subsequently, discriminative representations are acquired by minimizing the triplet loss, thereby enhancing the model's generalization capabilities. Extensive evaluations, including cross-working condition and cross-machine tasks, demonstrate the effectiveness and superiority of the proposed method.

Enhancing the Classification of Imbalanced Datasets through the Utilization of the Provisional Clarifier-Alternator (cPAen)

Warmcomfort data is vital for enhancing heating and cooling system efficiency through ML (ML) models. Yet, these datasets often suffer from severe class imbalance due to subjective feedback. To tackle this issue, we introduce a data augmentation algorithm, Provisional Clarifier-Alternator (cPAen). In our research, we evaluated cPAen using real Warmcomfort data and found it outperformed SMOTE, ADASYN, and cWGAN-GP in terms of F1 scores. Notably, cPAen is over 5 times faster than cWGAN-GP while maintaining high test accuracy.

Imbalanced learning for wind turbine blade icing detection via spatio-temporal attention model with a self-adaptive weight loss function

Accurate blade icing detection of wind turbines is of great significance to avoid secondary damages or accidents. Recently, data-driven methods have attracted increasing attention due to the availability of supervisory control and data acquisition (SCADA) data with their low-cost and high volume. However, SCADA data usually present complex characteristics of high dimensionality and strong spatio-temporal correlations among different sensor variables and severe class imbalance. It is challenging to extract effective features for accurate and reliable icing detection from imbalanced SCADA data. To address these challenges, we propose an imbalanced learning method for blade icing detection. First, we propose a spatio-temporal attention model to extract and assign weights to temporal-spatial features based on their contributions to blade icing detection. Then, to address the class imbalance issue, we design a self-adaptive weight (SAW) loss function to divide SCADA data into multiple batches and then classify them during model training. Unlike conventional loss functions, our SAW loss can adaptively assign weights to each data category according to their numbers in the divided batches. The effectiveness of the proposed method is verified on real SCADA datasets. The results demonstrate our designed spatio-temporal attention and SAW loss achieved better classification results and superior self-adaptability.

Analysis of South Korean Class Conflict Revealed by Elements of Lies and Violence in Cinema - Parasite as an Example

Since the implementation of neoliberal economic policies in South Korea in the 1990s, the huge gap between the rich and the poor has become a phenomenon, an acute social issue, and a frequent topic of film and television series, which has stimulated a series of topics and discussions critical of capitalism and anti-neoliberalism. This paper examines the class conflicts that exist in South Korean society in the context of the neoliberal economic system. It explores the representation of class conflict and the gap between rich and poor in South Korean film and television productions. This paper applies a case study approach to the film Parasite, through reviewing its narratives and the analysis of mise-en-scne, it deals with the elements of lies and violence to illustrate the class conflicts in South Korean society. The findings of this study highlight the presence of a severe class conflict in South Korean society, which is reflected in the immoral behavior of the lower-class characters in Parasite, as well as in the violent declaration of war within the lower class and towards the upper class.

Depressive Symptoms in Young and Middle-Aged Stroke Patients: A Transition Analysis.

There is heterogeneity in depressive symptoms. However, latent classes of depressive symptoms and the transition and influences of these in young and middle-aged stroke patients are unclear. The aim of this study was to identify the latent classes of depressive symptoms and their transition patterns over time and the influencing factors in young and middle-aged stroke patients from stabilization to 6 months after discharge. This is a longitudinal study following the Strengthening the Reporting of Observational Studies in Epidemiology checklist. A total of 272 young and middle-aged stroke participants were recruited from a hospital neurology ward in Henan Province, China. Participants completed a questionnaire on sociodemographic and health information. Latent transition analysis was used to evaluate the transition pattern of latent classes from stabilization to 6 months after discharge and its influencing factors. One hundred seventy-nine participants were included in the analysis. Three latent classes of depressive symptoms were identified as "mild symptoms," "grief-sleep-fatigue symptoms," and "severe symptoms." Most participants remained in the original latent class from stabilization to 6 months after discharge (probability of 83.8%, 83.8%, and 88.8%). From 3 to 6 months after discharge, the participants with fewer complications were more likely to transition into the mild symptom class. The findings indicate that from stabilization to 6 months after discharge, depressive symptoms in young and middle-aged stroke patients in China transitioned gradually from the severe symptom class to the mild symptom. Patients with fewer numbers of poststroke complications were more likely to transition to the mild symptoms class. Future research should focus on depressive symptoms in early-stage stroke patients and provide sufficient psychological support to patients with a high number of complications.

Multidisciplinary orthodontic-orthognathic management of severe skeletal Class III malocclusion and anterior open bite

This case report presents the multidisciplinary management of a patient with severe skeletal Class III malocclusion, anterior open bite, dentigerous cyst, and multiple restorations. The patient underwent orthodonticorthognathic treatment, autotransplantation, cyst removal, and esthetic rehabilitation with porcelain veneers. The presurgical orthodontic decompensation of the proclined upper incisors involved mini-screw-assisted entire upper arch distalization. The bimaxillary osteotomy included advancing the maxilla, mandibular setback, rotating the maxillomandibular complex clockwise, and moving the chin posteriorly. The role of multidisciplinary treatment is very important in orthognathic patients with compromised dentition. Autotransplantation could help patients avoid the need for prosthetic implants. Entire arch distalization could be a possible alternative to premolar extraction for presurgical orthodontic decompensation in cases of proclined incisors.

Evaluation of the mandibular condyles trabecular structure in patients with severe class III pattern: a computed tomography (CT) fractal analysis study

Class III malocclusion is a combination of dental and skeletal disorders that causes discrepancies in occlusion. Malocclusion can affect the structure of the Temporomandibular joint (TMJ) resulting in many problems, one of which is affecting the internal structure of the bone. This study aimed to examine the morphological features of class III patients’ trabecular structure of the mandibular condyle in comparison with normal class I individuals using fractal analysis (FA). To study the internal structure of the mandibular condyle bone, Computed Tomography (CT) scans of 45 severe class III patients (age: 16–46) who were the candidates for orthognathic surgery were selected and matched by age and sex with 45 normal class I individuals. The trabecular bone structure of the left and right mandibular condyles in three CT planes of the study group and control group were evaluated employing the FA. The result of the present study indicated that the fractal dimensions (FD) values of class III patients were lower than those of the normal class I individuals in axial (class I: 1.31 ± 0.02, class III: 1.28 ± 0.02), sagittal (class I: 1.25 ± 0.03, class III: 1.19 ± 0.08), and coronal (class I: 1.5 ± 0.06, class III: 1.45 ± 0.07) planes (P < 0.001). There were no significant differences between the FD values of the males and females. The intra-group evaluation revealed that there was no correlation between age and FD values. No evidence of laterality was found regarding the FD values of the right and left condyles. Given the noticeable differences between the FD values, it can be implied that severe class III malocclusion may affect the trabecular pattern of the cancellous bone of the mandibular condyle compared to class I individuals. Therefore, due to the altered trabecular structure, clinicians should be cautious when planning treatments for these patients.

Clasificación de gravedad y variables de influencia del síndrome poscuidado intensivo

ObjectiveThe study aims to characterize postintensive care syndrome by classifying the severity of the disease and identifying the variables of influence in 2 highly complex intensive care units for adults in Colombia. MethodA descriptive, cross-sectional, prospective study was carried out to characterize survivors of critical illness using the Healthy Aging Brain Care–Monitor in a sample of 135 patients. Postintensive care syndrome severity was classified using Gaussian mixture models for clustering, and the most influencing variables were identified through ordinal logistic regression. ResultsClustering based on Gaussian mixture models allowed the classification of postintensive care syndrome severity into mild, moderate, and severe classes, with an Akaike information criterion of 308 and an area under the curve of 0.80, which indicates a good fit; thus, the mild class was characterized by a score on the HABC-M Total scale≤9; the moderate class for a HABC-M Total score≥10 and≤42 and the severe class for a HABC-M Total score≥43. Regarding the most influencing variables, the probability of belonging to the moderate or severe classes was related to male sex (91%), APACHE II score (22.5%), age (13%), intensive care units days of stay (10.6%), the use of sedation, analgesia and neuromuscular blockers. ConclusionIntensive care units survivors were characterized using the HABC-M scale, which made it possible to classify postintensive care syndrome through Gaussian mixture models clustering into mild, moderate, and severe, and to identify variables that had the major influence on the presentation of postintensive care syndrome.

Soil erosion rates evaluation via 137Cs and RUSLE modeling in the Sigilil watershed (northeastern Tunisia)

Soil erosion is a natural phenomenon which causes many problems such as the reduction of soil fertility and silting of dams. In this study, the 137Cs activities and the RUSLE model were used to quantify soil erosion rates in the Sgilil river watershed of northeastern Tunisia, and where 20% of the agricultural land is devoted to the cultivation of olive trees. In addition, the erosion dynamics and its spatial variation were inferred. The net erosion rates in the cultivated field based on the measured 137Cs activities and using the Fallout Radionuclide Conversion Models range from 2.24 to 34.23 t ha−1 yr−1 when the Proportional Model is applied and from 1.49 to 30 t ha−1 yr−1 when the Mass Balance Model 2 is applied. The average erosion rates of the Sgilil river watershed estimated by the RUSLE model after multiplying five erosion factors for an average year (1990–2019) is 6 t ha−1 yr−1. Based on the RUSLE model results, around 52% of the watershed area falls within the slight erosion risk class (0–2 t ha−1 yr−1) and 11% within the very severe erosion risk class (> 30 t ha−1 yr−1). The center of the watershed, where the olive cultivated field is located belongs to a high to severe erosion risk class. Results attained through the isotopic approach and those found through the RUSLE modeling are in agreement.

Orthodontic Surgical Approach of Severe Skeletal Hyperdivergent Class II Malocclusion Treated by Le-Fort I Osteotomy, BSSO and Genioplasty

Orthodontic Surgical Approach of Severe Skeletal Hyperdivergent Class II Malocclusion Treated by Le-Fort I Osteotomy, BSSO and Genioplasty

Ön Eğitimli Modeller ve Özellik Seçiminin Rolü: Diyabetik Retinopati Tanısında Yapay Zeka Tabanlı Yaklaşım

Diabetic retinopathy is a significant complication occurring in the retina of the eye as a result of prolonged diabetes. When not detected early, this condition can lead to vision loss. Advanced image processing techniques and artificial intelligence algorithms have enhanced the possibilities of early diagnosis and treatment. This article discusses current advancements in artificial intelligence-based diabetic retinopathy detection and explores future possibilities in this field. In the experimental studies of the article, the Kaggle Aptos 2019 dataset was utilized. This dataset comprises 5 classes and a total of 3662 images. The class distribution is as follows: No DR (No Diabetic Retinopathy): 1805, Mild: 370, Moderate: 999, Severe: 193, Proliferative DR: 295. The study consists of four fundamental stages. These stages are (1) Feature extraction from VGG16 and VGG19 pretrained models, (2) Feature selection using NCA, Relieff, and Chi2, (3) Classification with Support Vector Machine classifier, (4) Iterative Majority Voting. Using the proposed method, a high accuracy of 99.18% is achieved. Furthermore, sensitivity of 100% for the No DR class, sensitivity of 100% for the Moderate class, sensitivity of 98.80% for the Severe class, and an F1-Score of 99.89% for the No DR class are obtained. This study demonstrates the effective utilization of machine learning methods in diabetic retinopathy diagnosis. The experimental results underscore the significant contributions of diabetic retinopathy patients' diagnosis and treatment processes.