Subjective Classification Research Articles

Classification of Rice Grains by Image Processing

General Background: The assessment of the surface quality of pre-treated rice grains is crucial for determining their market acceptability, storage stability, processing quality, and overall consumer satisfaction. Traditional evaluation methods are often time-consuming and yield subjective classifications. Specific Background: The lack of a comprehensive tagged image dataset hinders the application of advanced convolutional neural networks (CNN) for detailed damage classification of healthy rice grains. Knowledge Gap: Existing datasets and methods limit the effective exploration of sophisticated CNN models for categorizing rice types, particularly in identifying subtle damage characteristics. Aims: The study aims to create a robust rice grain classification system using image processing techniques, primarily deep learning algorithms, to improve the classification of rice varieties. Results: Utilizing a dataset of 75,000 images across five widely cultivated rice varieties in Turkey, we achieved classification accuracies of 100% for CNN, 99.95% for deep neural networks (DNN), and 99.87% for artificial neural networks (ANN). Novelty: The proposed approach represents a significant advancement in rice classification technology, employing a combination of image acquisition, feature extraction, and machine learning to streamline the process, effectively addressing the challenges faced in traditional methods. Implications: The findings underscore the potential for improved sorting and grading efficiency in the rice industry, facilitating better market outcomes and consumer satisfaction through enhanced quality control Highlights: Innovative: Uses CNN for accurate rice variety classification. Data-Driven: Analyzes 75,000 images for enhanced quality evaluation. Impactful: Increases efficiency in sorting and grading processes. Keywords: Rice Classification, Image Processing, CNN, Machine Learning, Agricultural Technology

Effect of acupuncture on the sleep quality of hemodialysis patients in a capital city in the Northeast of Brazil

Objective: Analyze the effect of auricular acupuncture on the sleep quality of hemodialytic patients with chronic pain in a capital city in the Northeast of Brazil. Methods: This study includes double-blind, randomized in blocks and allocated at a rate of 1:1, controlled trial that was conducted in two hemodialysis clinics, with 94 chronic renal patients with chronic pain (auricular acupuncture [AA]: 47, sham acupuncture [SA]: 47). Both received 12 sessions for three months. We collected data using the Pittsburg Sleep Quality Index (PSQI). Data were analyzed using t-tests, Wilcoxon test and ANOVA. Results: Comparing the SA and AA groups, there was a statistically significant difference at baseline (p < 0.001), a significant intergroup difference, and the AA group had better sleep quality than the SA group during and at the end of the 12 sessions (< 0.001). A high magnitude of the effect of auricular acupuncture on sleep quality was found at the end of the sessions (D= -1.94). Conclusion: The auricular acupuncture intervention proved to be effective in improving the subjective sleep classification of renal patients.

Classifying fabric wrinkle via regularized random vector function link optimized by African vultures optimization algorithm

The wrinkling of fabrics not only affects their aesthetic appeal and comfort but also leads to increased wear and a reduced lifespan. Consequently, wrinkle resistance has become a critical metric in evaluating fabric performance. Traditional wrinkle grading methods rely heavily on expert assessments, which are susceptible to visual fatigue, resulting in subjective and inconsistent classification outcomes. To address this issue, this article proposes an automatic wrinkle grading model based on an improved African Vultures Optimization (AVO) algorithm, which optimizes a regularized random vector functional link (RRVFL) network. The model enhances convergence speed and classification accuracy by incorporating the Harris Hawks Optimization (HHO) algorithm to provide a superior initial population for the AVO algorithm. Additionally, the AVO algorithm’s local search capability is strengthened during the optimization process, improving the model’s stability and resistance to overfitting. Experimental results demonstrate that the proposed HHO-AVO-RRVFL model achieves an average classification accuracy of 97.86% across multiple datasets, significantly outperforming other comparative algorithms while also excelling in classification stability and convergence speed.

Subjective well-being of college students: Developmental trajectories, predictors, and risk for depression

The study aimed to identify different developmental trajectories of subjective well-being for 2 473 college students from freshman to senior year (female = 47.15%; mean age = 19.6 years, SD = 0.894 years). We further explored the associations between categories of well-being development and depression levels across different groups of subjective well-being. Students completed a longitudinal survey over four academic years, self-reporting their subjective well-being, mental state, demographic characteristics, family background, and academic performance. Growth mixture model analysis revealed three distinct developmental trajectories of college students’ subjective well-being: high stable (HS), low increasing (LI), and high decreasing (HD). Factors such as sex, physical health, self-esteem, parental education level, and peer relationships were significantly associated with the subjective well-being classification trajectories. In addition, a stable and high level of subjective well-being was associated with lower risk for depression. The results indicate the importance of subjective well-being practices for students’ long-term schooling-related quality of life.

Impact of Effective Word Vectors on Deep Learning Based Subjective Classification of Online Reviews

Sentiment Analysis tasks are made considerably simpler by extracting subjective statements from online reviews, thereby reducing the overhead of the classifiers. The review dataset encompasses both subjective and objective sentences, where subjective writing expresses the author's opinions, and objective text presents factual information. Assessing the subjectivity of review statements involves categorizing them as objective or subjective. The effectiveness of word vectors plays a crucial role in this process, as they capture the semantics and contextual cues of a subjective language. This study investigates the significance of employing sophisticated word vector representations to enhance the detection of subjective reviews. Several methodologies for generating word vectors have been investigated, encompassing both conventional approaches, such as Word2Vec and Global Vectors for word representation, and recent innovations, such as like Bidirectional Encoder Representations from Transformers (BERT), ALBERT, and Embeddings from Language Models. These neural word embeddings were applied using Keras and Scikit-Learn. The analysis focuses on Cornell subjectivity review data within the restaurant domain, and metrics evaluating performance, such as accuracy, F1-score, recall, and precision, are assessed on a dataset containing subjective reviews. A wide range of conventional vector models and deep learning-based word embeddings are utilized for subjective review classification, frequently in combination with deep learning architectures like Long Short-Term Memory (LSTM). Notably, pre-trained BERT-base word embeddings exhibited exceptional accuracy of 96.4%, surpassing the performance of all other models considered in this study. It has been observed that BERT-base is expensive because of its larger structure.

Design of Music Emotion Analysis and Creation Aid System Based on Machine Learning

In human-centric contexts, visual vision plays an increasingly important role in decision making, learning, communication, and situation awareness. Music is a language that conveys emotion to all creatures, including plants and animals. These emotional characteristics are included in the emotion found in music, along with other feelings conveyed via melodies, rhythms, and composer-highlighted moments. Human interaction, communication, decision-making, and cognitive processes are all heavily influenced by emotion. However, Subjective music emotion classification is made more difficult by pure machine classification due to its objectivity; additionally, emotion classification has not reached the same level of accuracy as style classification due to the complex and poorly understood mechanisms underlying how humans perceive music and generate emotions. This manuscript proposes a Multi-Scale Adaptive Graph Neural Network (MSAGNN)optimized with the Remora Optimization Algorithm (ROA) for predicting music emotion MEA-MSAGNN-ROA. Initially data is taken from Music Video Emotion (MVE) dataset Afterward the data is fed to Attribute-based Neural Collaborative Filtering (AbNCF) based pre-processing process. The outcome from the pre-processing data is transferred to the MSAGNN. The music emotions are successfully classified by using MSAGNN are exciting, fear, sad and relaxation. The ROA is used to optimize the weight parameter of MSAGNN. The proposed MEA-MSAGNN-ROA is applied in Python working platform. Performance parameters, like accuracy, precision, sensitivity, F1-score, ROC and recall analyzed to compute proposed approach. The proposed MEA-MSAGNN-ROA approach yields improved results in terms of accuracy (22.46%, 38.58%, 21.74%), sensitivity (21.97%, 33.88%, 25.52%), and precision. reduced computation times of 21.86%, 36.76%, 28.95%, and 81.46%, 95.97%, 86.77%.The proposed MEA-MSAGNN-ROA method is compared with the existing methods such as MEA-CNN, MEA-BPNN, and MEA-FNN, respectively. From the result it is concludes that the proposed MEA-MSAGNN-ROA method based accuracy is higher than the existing methods.

Insights into Psychoactive Drug Effects: The Role of Drug Discrimination Techniques

Psychoactive drugs exert profound effects on the central nervous system, influencing behavior and perception. Understanding these effects is crucial for both scientific research and clinical practice. This article provides an overview of the application of drug discrimination techniques in studying the subjective effects, pharmacological mechanisms, individual differences, and drug classification of psychoactive substances. Through drug discrimination models, researchers gain insights into the pharmacological specificity and predictive validity of these substances, particularly hallucinogens such as LSD and cannabinoids like THC. The article highlights the contributions of drug discrimination techniques in elucidating the pharmacological actions of psychoactive drugs, offering opportunities for further exploration into their therapeutic potential and abuse-related properties.

Optimal cutoff value of the dry eye-related quality-of-life score for diagnosing dry eye disease

This retrospective study aimed to determine the optimal cutoff values of the Dry Eye-Related Quality-of-Life Score (DEQS) questionnaire for diagnosing dry eye disease (DED) and classifying DED severities. Participants completed the DEQS questionnaire, the Japanese version of the Ocular Surface Disease Index (J-OSDI) questionnaire, and DED examinations. DED was diagnosed according to the 2016 Asia Dry Eye Society diagnostic criteria based on DED symptoms (J-OSDI ≥ 13 points) and tear film breakup time ≤ 5 s. Receiver operating characteristic (ROC) analysis was used to calculate the optimal cutoff values of the DEQS summary score for detecting DED and grading its severity. Among 427 patients, 296 (69.3%) and 131 (30.7%) were diagnosed with DED and non-DED, respectively. ROC analysis determined an optimal cutoff value of 15.0 points for DED diagnosis, with 83.5% sensitivity, 87.0% specificity, and an area under the curve of 0.915. The positive and negative predictive values for DEQS ≥ 15.0 points were 93.6% and 69.9%, respectively. DEQS cutoff values of 15.0, 20.0, and 26.8 points could be accepted for severity classification of DED subjective symptoms in clinical use and represent mild, moderate, and severe DED, respectively. Conclusively, the optimal cutoff values of DEQS enable DED detection and subjective symptom severity classification.

Identification of residential building typologies by applying clustering techniques to cadastral data

Building typologies are usually classified according to their shape, distribution and construction features depending on the time period they were built. As a result, a subjective classification arises which highly depends on the criteria used to differentiate buildings. When this analysis is carried out at the urban scale, data sets become bigger, making patterns difficult to uncover. Mistakes in deciding the variables to classify buildings lead to incorrect typologies and, consequently, wrong results. The aim of this study is to test a new methodology based on clustering techniques to identify typologies related to energy retrofitting, which would allow obtaining a more objective classification and better pattern recognition by reducing human intervention. To that end, a data set from the Spanish cadastre is used, with additional information to reflect the influence of existing standards on constructive solutions. By applying three clustering techniques (Ward's method, Partitioning Around Medoids, and a combination of both), new proposals of building typologies are obtained and discussed in comparison to traditional classifications. The results show that the Ward's method produces building typologies with significantly high quality metrics and meaningfulness. The agglomeration coefficient is 99.8%, which indicates that hardly another hierarchical method could generate a better clusters structure. Six clusters comprise 86% of the dwellings as most occupants do not declare retrofitting works, thus not being reflected in the cadastre database. This research provides a new classification method that can notably influence the estimation of costs, environmental impact and cost effectiveness of energy retrofitting actions at urban scale.

Correlation between pathology and quantitative ultrasonographic evaluation of pancreatic fat with ultrasonographic classification.

No studies of the relationship between grayscale sonographic findings and pancreatic fat content have been reported to date. This study aimed to investigate the correlation between echogenicity and fat content of resected specimens using quantitative analysis. Forty-two consecutive patients who underwent pancreatoduodenectomy or distal pancreatectomy for pancreatic tumors were enrolled in this study. Ultrasonographic images were compared with quantitative pathological analysis. Subjective evaluation of echogenicity was classified as hypoechoic, isoechoic, hyperechoic, and super hyperechoic. The total and intralobular fat areas were measured. The mean, median, modal, minimum, and maximum ultrasound gray values correlated with the proportion of total fat area (r = 0.349; 0.357, 0.486, 0.466, and 0.347; p = 0.024, 0.020, 0.014, 0.019, and 0.089, respectively), but did not correlate with the proportion of intralobular fat area. Subjective classification was correlated with median gray value (p < 0.001), intralobular fat area (p = 0.118), and total fat area (p = 0.011). Cases were classified as hypoechoic (n = 3), isoechoic (n = 7), hyperechoic (n = 30), and super hyperechoic (n = 2). The subjective classification was correlated with the median gray value (p < 0.001) and total fat area (p = 0.005), and not correlated with the intralobular fat area (p = 0.118). Hyperechoic or super hyperechoic pancreatic parenchyma contains over 19.7% fat. Computed tomography values correlated with the proportion of intralobular fat area (r = -0.479, p = 0.004) and total fat area (r = -0.541, p < 0.001). Echogenicity classified based on subjective evaluation and image analysis were correlated with the proportion of fat in the pancreas.

The Urgency of Maqasid Shariah in Combating Religious Extremism from Ibn Ashur's Perspective

Religious extremism has become a prominent issue widely discussed worldwide today. Numerous studies from various perspectives have been conducted by scholars from both the Western and Eastern traditions to understand and identify solutions to the problem. Although the issue of religious extremism is considered somewhat sensitive due to its subjective classification, sometimes based on the biases of the evaluators, the Islamic perspective on progressive fiqh issues needs to be analyzed and highlighted as one of the efforts to address the phenomenon of extremism. This aims to prevent greater harm to the Muslim community and, consequently, preserve religion, life, intellect, honor, and wealth as desired by Islamic law (maqasid Sharia). This article employs a qualitative approach using a descriptive-analytical method, which seeks to critically describe the substance of various concepts and theories related to the moderate concept according to Ibn Ashur. The primary data sources are Ibn Ashur's written works discussing the Maqasid Sharia. Additionally, literature related to religious extremism will be utilized. This article is expected to contribute to the understanding and handling of religious extremism by integrating the principles of Maqasid Sharia from Ibn Ashur's perspective. The purpose of this article is to identify and analyze the significance of Maqasid Sharia in the context of combating religious extremism based on Ibn Ashur's perspective. It also aims to provide new insights into efforts to combat religious extremism by detailing the importance of Maqasid Sharia according to Ibn Ashur's thoughts.

A novel production profile classification system for incoming calves that predicts feedlot growth performance

Context Mitigating financial risk in the feedlot environment is an ongoing occurrence, and good production is a key risk mitigator. However, production protocols are based on historic averages because of the inability to predict growth potential of incoming calves. Production profiling of individual incoming feeder calves could address these limitations. Aims The aim of this study was to establish criteria for optimal sorting of incoming feeder calves into various cattle groups in a feedlot that maximises feedlot profit. Methods South African feeder calves (n = 436) were classified into four production-profile (PP) categories according to a predetermined set of phenotypic traits: PP 3 (n = 72) representing feeder calves with the poorest feedlot growth potential, PP 2− (n = 191) with below-average potential, PP 2+ (n = 139) with above-average potential and PP 1 (n = 34) with above-average feedlot growth potential. After combining the data of PP 2− and PP 2+ into PP 2, mixed modelling of economically important feedlot growth traits (average daily gain (ADG), carcass ADG, and carcass exit weight) was performed to evaluate the effect of PP classification (PP 1 and PP 3), while adjusting for potential confounding effects such as starting weight (entry weight) and gender. Key results Carcass weights for calves with a PP classification of 3 and 1 were 15.54 kg less (P &lt; 0.000), and 11.34 kg more (P = 0.007) respectively, than those with a PP classification of 2 (261.27 kg, 95% CI 257.94–264.57), after adjusting for entry weight, calf gender and the random effect of the feeding pen. Similar to carcass weight, calves with a PP 3 classification were outperformed by other classifications in all the measured traits (P &lt; 0.05). Conclusions This is the first report demonstrating the ability of subjective production-profile classification to predict growth performance of individual feeder calves. Implications The opportunity of the PP classification system lies in value-based procurement of incoming feeder calves based on their growth potential at the start of the feeding period, and then to use technology to improve and finalise the current subjective PP classification system.

Development of Auto Dense-breast Classification on Mammography Images Using Image Similarity

In Japan, radiologists perform qualitative visual classification to define four categories of mammary gland density. However, an objective estimation of mammary gland density is necessary. To address this, we developed an automatic classification software using image similarity. We prepared 741 cases of mediolateral oblique images (MLO) for evaluation, and they were diagnosed as normal among the mammography images taken at our hospital. Image matching was performed using the evaluation images and an image database for breast density determination. In this study, the image similarity used zero normalized cross-correlation (ZNCC) as an index. In addition, if the breast thickness is less than 30 mm and each breast density category ZNNC has the same value, the category is evaluated on the fat side. We compared the results of qualitative visual classification and automatic classification methods to assess consistency. The agreement with the subjective breast composition classification was 78.5%, and the weighted kappa coefficient was 0.98. One mismatched case was evaluated on the higher density side with the same ZNCC value between categories and a breast thickness greater than 30 mm. Image similarity provides an excellent estimation of quantification of breast density. This system could contribute to improving the efficiency of the mammography screening system.

Utility of texture analysis for objective quantitative ex vivo assessment of meningioma consistency: method proposal and validation.

Tumor consistency is considered to be a critical factor for the surgical removal of meningiomas and its preoperative assessment is intensively studied. A significant drawback in the research of predictive methods is the lack of a clear shared definition of tumor consistency, with most authors resorting to subjective binary classification labeling the samples as "soft" and "hard." This classification is highly observer-dependent and its discrete nature fails to capture the fine nuances in tumor consistency. To compensate for these shortcomings, we examined the utility of texture analysis to provide an objective observer-independent continuous measure of meningioma consistency. A total of 169 texturometric measurements were conductedusing the Brookfield CT3 Texture Analyzer on meningioma samples from five patients immediately after the removal and on the first, second, and seventh postoperative day. The relationship between measured stiffness andtime from sample extraction, subjectively assessedconsistency grade and histopathological features (amountof collagen and reticulin fibers, presence of psammoma bodies, predominant microscopicmorphology) wasanalyzed. The stiffness measurements exhibited significantly lower variance within a sample than among samples (p = 0.0225) and significant increase with a higher objectively assessed consistency grade (p = 0.0161, p = 0.0055). A significant negative correlation was found between the measured stiffness and the time from sample extraction (p < 0.01). A significant monotonic relationship was revealed between stiffness values and amount of collagen I and reticulin fibers; there were no statistically significant differences between histological phenotypes in regard to presence of psammoma bodies and predominant microscopic morphology. We conclude that the values yielded by texture analysis are highly representative of an intrinsic consistency-related quality of the sample despite the influence of intra-sample heterogeneity and that our proposed method can be used to conduct quantitative studies on the role of meningioma consistency.

MRI-Based Brain Tumor Classification Using Inception Resnet V2

Brain tumors are one of the most fatal disorders owing to the uncontrolled proliferation of abnormal cells inside the brain. Digital images are obtained using Magnetic Resonance Imaging (MRI), which is a medical instrument that can assist doctors and other medical personnel in assessing and diagnosing the presence and type of brain tumors. However, manual and subjective classification is time-consuming and error prone. Hence, an objective, automatic, and more reliable method is needed to classify MRI images of brain tumors. Artificial intelligence is considered appropriate to determine the type of brain tumor via MRI images to overcome the constraints of conventional testing methods. One method for performing automatic classification is the Convolutional Neural Network (CNN). This work demonstrates how the Inception Resnet v2 architecture in CNN is utilized to classify MRI brain tumors into four categories via transfer learning, namely glioma tumors, meningioma tumors, no tumors, and pituitary tumors. The accuracy value of the generated model reached 93.4% after running for 20 epochs. It infers that artificial intelligence is beneficial in identifying a brain tumor objectively to help doctors and radiologists in the medical field.

Time to change? Present and prospects of hemorrhoidal classification.

As a common benign anal condition, the high incidence and recurrence of hemorrhoids pose challenges for both patients and doctors. The classification of hemorrhoids plays a crucial role in assessing, diagnosing, and treating the condition. By using appropriate classification and corresponding treatment strategies, we can achieve higher cure rates and lower recurrence rates of hemorrhoids. Since the introduction of the Miles classification in 1919, various classifications have been developed, which include objective classifications based on anatomical or instrumental assessment and subjective classifications based on symptoms and patient sensations. These classifications aim to accurately evaluate the condition. In this study, we discuss the evaluation values of each classification in terms of their advantages, disadvantages, treatment relevance, reproducibility, practicality, and assessment value. We also analyze the significant and essential factors, principles of use, and components of assessment indicators of hemorrhoidal classification. This study proposes several strategies to address the limitations of current hemorrhoidal assessment methods. All these will provide a reference for the development regarding the assessment and classification of hemorrhoids and clinical diagnosis and management of hemorrhoids.

Classification of COVID-19 X-ray images using transfer learning with visual geometrical groups and novel sequential convolutional neural networks

COVID-19 is a highly transmissible infectious disease that remains a substantial challenge. The utilization of chest radiology, particularly X-ray imaging, has proven to be highly effective, easily accessible, and cost-efficient in detecting COVID-19. A dataset named COVID-Xray-5k, consisting of imbalanced X-ray images of COVID-19-positive and normal subjects, is employed for investigation. The research introduces a novel methodology that utilizes conventional machine learning (ML), such as local binary patterns (LBP) for feature extraction and support vector machines (SVM) for classification. In addition, transfer learning is employed with the Visual Geometry Group 16-layer (VGG16) and 19-layer (VGG19) models. Besides, novel sequential convolutional neural network (CNN) architectures are presented to develop an autonomous system for classifying COVID-19. One of the proposed CNN architectures classifies the test dataset with an F1 score of 91.00% and an accuracy of 99.45% based on an empirical investigation to determine optimal hyper-parameters. The methods presented in the research show promising potential for COVID-19 classification, irrespective of class imbalance.•Employment of ML models to investigate subjective feature engineering and classification.•Transfer learning was employed for VGG16 and VGG19 with eight distinct models.•Illustration of two novel CNN sequential architectures; all the investigation is performed with and without weighted sampling.

P-757 The classification and management of adnexal masses identified during pregnancy

Abstract Study question How do we classify adnexal masses (ovarian cysts) identified during pregnancy and how should we manage them? Summary answer Adnexal masses (ovarian cysts) should be managed expectantly in pregnancy, given the low rate of complications and low prevalence of malignancy. What is known already Conservative management of adnexal masses (ovarian cysts) is preferable during pregnancy, due to the maternal and fetal risks associated with surgery. Accurate diagnosis is key to safe expectant management. Various ultrasound-based models exist to classify adnexal masses; however, none have been validated for use in pregnancy. Adnexal mass morphology can change during pregnancy due to decidualisation, which is often difficult to distinguish from underlying neoplastic processes. We aim to evaluate:(1) the performance of current methods of adnexal mass classification in pregnancy, (2) understand the natural course of adnexal masses in pregnancy, particularly the presence of decidualisation and incidence of complications. Study design, size, duration Retrospective analysis of prospectively collected data between January 2017- November 2022. To classify adnexal masses (ovarian cysts), we evaluated: Expert subjective assessment (SA), IOTA Simple Rules (SR) and the IOTA Assessment of Different Neoplasias of the Adnexa (ADNEX) model. The end point was either the histological examination of tissue removed at surgery or the subjective classification of adnexal masses at the postnatal scan in women managed conservatively. Participants/materials, setting, methods Women with an adnexal mass identified on gynaecological ultrasound in pregnancy at a tertiary London University Hospital were included. Relevant clinical data was extracted, including age, gestation and cyst-related complications. Adnexal masses were classified at the first antenatal and postnatal ultrasound examination according to SA, SR and the ADNEX model (10% risk of malignancy cut off) and correlated to histology for adnexal masses managed surgically. IOTA simple descriptors were used to classify benign adnexal masses. Main results and the role of chance 254 women (median age 33-years old, range:18-49) with an adnexal mass/cyst were included at a median gestation of 12 weeks, (range:4-36). 13 (5.1%) conceived through assisted reproductive techniques. Spontaneous resolution occurred in 24.5% of cases, and 21 were lost to follow up (8.3%). According to Simple Descriptors, 38.6% were simple, 22.0% were endometriomas and 19.7% were dermoid cysts. Antenatally, SA outperformed ADNEX, based on specificity (94.6 vs.92.2%) and sensitivity (60 vs.55.6%) respectively. SA and ADNEX had a negative predictive value(NPV) of 96.1% and 98.3% respectively. Postnatally, SA had a higher sensitivity than ADNEX (75 vs.50%), but a lower specificity (94.1 vs.96.9%). SA and ADNEX had a NPV of 94.1% and 97.5% respectively. Three (1.3%) underwent acute surgery during pregnancy: one ovarian torsion (9/40), and two cyst ruptures (11/40, 30/40). Presumed decidualisation occurred in 29.5% of endometriomas. Three (1.3%) underwent cyst removal during caesarean section, two for suspicion of BOTs (one had suspected decidualisation) and one for patient preference. On histology, the suspected decidualised mass was a serous BOT, and the suspected BOT was a struma ovarii. In the postnatal period (12 weeks), four underwent surgery for suspicion of BOTs, two were confirmed BOTs and two were benign cystadenomas on histology. Limitations, reasons for caution The dataset as expected contains a relatively small number of malignancies (prevalence 3.9%) and so any analysis of test performance to discriminate between benign and malignant disease must be interpreted with caution. A larger prospective multicentre study is needed to do this. Wider implications of the findings Our data suggest that a number of cysts in pregnancy will resolve, a few will be associated with acute complications and the risk of malignancy is very low. Accordingly, the study supports an expectant management approach for adnexal masses (ovarian cysts) detected in pregnancy. Trial registration number Not applicable

A deep learning model enables accurate prediction and quantification of pulmonary edema from chest X-rays

BackgroundA quantitative assessment of pulmonary edema is important because the clinical severity can range from mild impairment to life threatening. A quantitative surrogate measure, although invasive, for pulmonary edema is the extravascular lung water index (EVLWI) extracted from the transpulmonary thermodilution (TPTD). Severity of edema from chest X-rays, to date is based on the subjective classification of radiologists. In this work, we use machine learning to quantitatively predict the severity of pulmonary edema from chest radiography.MethodsWe retrospectively included 471 X-rays from 431 patients who underwent chest radiography and TPTD measurement within 24 h at our intensive care unit. The EVLWI extracted from the TPTD was used as a quantitative measure for pulmonary edema. We used a deep learning approach and binned the data into two, three, four and five classes increasing the resolution of the EVLWI prediction from the X-rays.ResultsThe accuracy, area under the receiver operating characteristic curve (AUROC) and Mathews correlation coefficient (MCC) in the binary classification models (EVLWI < 15, ≥ 15) were 0.93 (accuracy), 0.98 (AUROC) and 0.86(MCC). In the three multiclass models, the accuracy ranged between 0.90 and 0.95, the AUROC between 0.97 and 0.99 and the MCC between 0.86 and 0.92.ConclusionDeep learning can quantify pulmonary edema as measured by EVLWI with high accuracy.

Who is disabled? On whether the functional definition of disability targets the same individuals as the subjective definition

With the increased attention to disability as a vulnerability criterion in the Sustainable Development Goals, international organizations and NGOs within the international development sector have started to pay explicit attention to persons with disabilities, including the collection of data on persons with disabilities. The Washington Group Short Set of Questions, which focuses on functional limitations, has been gaining popularity as an assessment tool for disability. This set of questions reflects a categorization of disability that does not necessarily correspond with subjective disability assessments, such as the yes/no question (“do you have a disability?”) which many development actors have used in their assessment tools when they collect disability data This study compares the subjective and the functional limitations assessment tools for disability to answer the question: do they identify the same individuals as persons with disabilities? Based on a survey carried out amongst persons with disabilities in Cambodia, we included both the Washington Group Short Set and a subjective question asking respondents to self-identify their disability type. We find that, although all respondents self-identified as disabled, not all respondents would be considered disabled according to the Washington Group Short Set of questions. In addition, there is little overlap between specific disability types according to a subjective classification method and the domains of functioning measured through the Washington Group methodology. Our findings affirm that categorization as abled or disabled depends on the tool used. This is important, as the assessment approach chosen by those collecting disability data can shape the design choices of policies and programs, and determine who benefits.