Automatic Screening Research Articles

A portable multi-channel fluorescent paper-based microfluidic chip based on smartphone imaging for simultaneous detection of four heavy metals

Due to the excessive contamination of heavy metals pollution, it is very urgent and necessary to develop a real-time detection method for the heavy metals in food. As a target sensing device, a paper-based microfluidic device (μPAD) has the advantages of simplicity, low-cost, and portability. In this study, a self-driven microfluidic paper-based chip was first developed for the simultaneous detection of four targets. The channels on the microfluidic chip were prepared by using wax printing and automatic screen printing on the filter paper, where liquid flowed by capillary force without pump assistance. Based on the specific binding ability of aptamers to heavy metals, a “turn-on” fluorescence aptasensor for the simultaneous detection of four heavy metal ions was developed on the proposed multi-channel device via smartphone imaging. The obtained fluorescence images were digitized into RGB color values by Image J software, and an M-mode was established to realize the quantitative detection of heavy metal ions. Under optimal conditions, the limits of detection for lead(II), mercury(II), cadmium(II), and arsenic(III) were 4.20 nM, 1.70 nM, 2.04 nM, and 1.65 nM, respectively. Furthermore, the aptasensor was successfully applied to the quantitative detection of four heavy metal ions in apple and lettuce samples with recovery rates of 84.0%–104.1%.

Modeling new trends in bone regeneration, using the BERTopic approach.

Aim: Bibliometric surveys are time-consuming endeavors, which cannot be scaled up to meet the challenges of ever-expanding fields, such as bone regeneration. Artificial intelligence, however, can provide smart tools to screen massive amounts of literature, and we relied on this technology to automatically identify research topics. Materials & methods: We used the BERTopic algorithm to detect the topics in a corpus of MEDLINE manuscripts, mapping their similarities and highlighting research hotspots. Results: Using BERTopic, we identified 372 topics and were able to assess the growing importance of innovative and recent fields of investigation such as 3D printing and extracellular vescicles. Conclusion: BERTopic appears as a suitable tool to set up automatic screening routines to track the progress in bone regeneration.

A multicenter study on two-stage transfer learning model for duct-dependent CHDs screening in fetal echocardiography

Duct-dependent congenital heart diseases (CHDs) are a serious form of CHD with a low detection rate, especially in underdeveloped countries and areas. Although existing studies have developed models for fetal heart structure identification, there is a lack of comprehensive evaluation of the long axis of the aorta. In this study, a total of 6698 images and 48 videos are collected to develop and test a two-stage deep transfer learning model named DDCHD-DenseNet for screening critical duct-dependent CHDs. The model achieves a sensitivity of 0.973, 0.843, 0.769, and 0.759, and a specificity of 0.985, 0.967, 0.956, and 0.759, respectively, on the four multicenter test sets. It is expected to be employed as a potential automatic screening tool for hierarchical care and computer-aided diagnosis. Our two-stage strategy effectively improves the robustness of the model and can be extended to screen for other fetal heart development defects.

An approach of filtering simulated magnetospheric X-ray images based on self-supervised network and random forest

Imaging is an important method for observing the Earth’s space environment. Future missions, such as the Solar wind Magnetosphere Ionosphere Link Explorer (SMILE), aim to explore the interaction between the solar wind and the Earth’s magnetosphere via soft x-ray imaging. With the advent of these missions, a large number of magnetospheric images may be acquired. However, as the viewing geometry and solar wind conditions change, satellites sometimes fail to capture the magnetopause inside the field of view. We propose an approach that blends machine learning and deep learning to filter the simulated x-ray images for the SMILE mission, aiming to achieve automatic classification of the detected images. First, we performed magnetohydrodynamic simulations to derive the predicted SMILE x-ray images. Then, we used a self-supervised contrast feature extraction network to study the features of the images. Using this network, the random forest classifier can distinguish whether the subsolar point at the magnetopause has been detected. Finally, we designed the magnetopause filter to obtain the subsolar magnetopause images with observation positions outside the magnetosphere. As a result, the prediction accuracy of the classifier is up to 93%. And the F1 score is up to 95.5%. The stratified predictions allow an automatic screening of whether satellite magnetospheric images cover the subsolar magnetopause. These images, which have observation positions outside the magnetosphere, can be used to invert the three-dimensional magnetopause.

An Automatic and Objective Vision Screening System Based on EEG Signals

Vision screening cannot be popularized due to the uneven distribution of medical resources and subjective limitations of traditional detection. This paper proposes an automatic and objective vision screening system, which uses Electroencephalogram (EEG) signals to analyze people’s visual state, avoiding subjective errors. Meanwhile, this system just relies on personal operation without the involvement of medical staff, reducing the cost of time and medical resources. The experimental results show that the classification accuracy of the visual state can be increased to 80%, with the combined use of signal processing, relevant feature extraction, and optimization measures in this system. It indicates that the system is feasible and can play a role in visual screening based on EEG signals in the future.

Usefulness of UF-5000 automatic screening system in UTI diagnosis.

We have evaluated the Sysmex UF-5000 cytometer use in microbiology for the screening of negative urines, looking for cut-off points to detect bacteria and leukocytes. The number of processed urines was 3569, the highest to date in these studies. The best general cut-off point has been 100 bact/μl, giving an area under the ROC curve of 0.868, a sensitivity of 96%, a specificity of 50%, 1.17% of false negatives, and saving 40% of cultures. The PPV and NPV have been 35.5 and 95.4 respectively. The leukocyte count has not been useful. Finally, we have evaluated urine screening usefulness, concluding that in laboratories such as ours (284 urines/working day) or smaller, it is not cost-effective.

Automatic identification of medically important mosquitoes using embedded learning approach-based image-retrieval system

Mosquito-borne diseases such as dengue fever and malaria are the top 10 leading causes of death in low-income countries. Control measure for the mosquito population plays an essential role in the fight against the disease. Currently, several intervention strategies; chemical-, biological-, mechanical- and environmental methods remain under development and need further improvement in their effectiveness. Although, a conventional entomological surveillance, required a microscope and taxonomic key for identification by professionals, is a key strategy to evaluate the population growth of these mosquitoes, these techniques are tedious, time-consuming, labor-intensive, and reliant on skillful and well-trained personnel. Here, we proposed an automatic screening, namely the deep metric learning approach and its inference under the image-retrieval process with Euclidean distance-based similarity. We aimed to develop the optimized model to find suitable miners and suggested the robustness of the proposed model by evaluating it with unseen data under a 20-returned image system. During the model development, well-trained ResNet34 are outstanding and no performance difference when comparing five data miners that showed up to 98% in its precision even after testing the model with both image sources: stereomicroscope and mobile phone cameras. The robustness of the proposed—trained model was tested with secondary unseen data which showed different environmental factors such as lighting, image scales, background colors and zoom levels. Nevertheless, our proposed neural network still has great performance with greater than 95% for sensitivity and precision, respectively. Also, the area under the ROC curve given the learning system seems to be practical and empirical with its value greater than 0.960. The results of the study may be used by public health authorities to locate mosquito vectors nearby. If used in the field, our research tool in particular is believed to accurately represent a real-world scenario.

An artificial intelligence (AI)-based approach to clinical trial recruitment: The impact of Viz RECRUIT on enrollment in the EMBOLISE trial.

EMBOLISE (NCT04402632) is an ongoing randomized controlled trial investigating the safety and efficacy of middle meningeal artery embolization for the treatment of subacute or chronic subdural hematoma (SDH). Viz RECRUIT SDH is an artificial intelligence (AI)-based software platform that can automatically detect SDH in noncontrast computed tomography (NCHCT) images and report the volume, maximum thickness, and midline shift. We hypothesized that the mobile recruitment platform would aid enrollment and coordinate communication and image sharing among the entire research team. Patient enrollment in EMBOLISE prior to and after implementation of Viz RECRUIT SDH at a large comprehensive stroke center was compared along with the performance of the software platform. The EMBOLISE trial was activated on May 5, 2021, and Viz RECRUIT SDH was activated on October 6, 2021. The pre-AI cohort consisted of all patients from EMBOLISE to AI activation (153 days), and the post-AI cohort consisted of all patients from AI activation until August 18, 2022 (316 days). All alerts for suspected SDH candidates were manually reviewed to determine the positive predictive value (PPV) of the algorithm. Prior to AI-software implementation, there were 5 patients enrolled (0.99 patients/month) and one screen failure. After the implementation of the software, enrollment increased by 36% to 1.35 patients/month (14 total enrolled), and there were no screen failures. Over the entire post-AI period, a total of 6244 NCHCTs were processed by the system with 207 total SDH detections (3% prevalence). 35% of all alerts for suspected SDH were viewed within 10 min, and 50% were viewed within an hour. The PPV of the algorithm was 81.4 (CI [75.3, 86.7]). The implementation of an AI-based software for the automatic screening of SDH patients increased the enrollment rate in the EMBOLISE trial, and the software performed well in a real-world, clinical trial setting.

Automatic retinoblastoma screening and surveillance using deep learning.

Retinoblastoma is the most common intraocular malignancy in childhood. With the advanced management strategy, the globe salvage and overall survival have significantly improved, which proposes subsequent challenges regarding long-term surveillance and offspring screening. This study aimed to apply a deep learning algorithm to reduce the burden of follow-up and offspring screening. This cohort study includes retinoblastoma patients who visited Beijing Tongren Hospital fromMarch 2018 to January 2022 for deep learning algorism development. Clinical-suspected and treated retinoblastoma patients from February 2022 to June 2022 were prospectively collected for prospective validation. Images from the posterior pole and peripheral retina were collected, and reference standards were made according to the consensus of the multidisciplinary management team. A deep learning algorithm was trained to identify "normal fundus", "stable retinoblastoma" in which specific treatment is not required, and "active retinoblastoma" in which specific treatment is required. The performance of each classifier included sensitivity, specificity, accuracy, and cost-utility. A total of 36,623 images were included for developing the Deep Learning Assistant for Retinoblastoma Monitoring (DLA-RB) algorithm. In internal fivefold cross-validation, DLA-RB achieved an area under curve (AUC) of 0.998 (95% confidence interval [CI] 0.986-1.000) in distinguishing normal fundus and active retinoblastoma, and 0.940 (95%CI 0.851-0.996) in distinguishing stable and active retinoblastoma. From February 2022 to June 2022, 139 eyes of 103 patients were prospectively collected. In identifying active retinoblastoma tumours from all clinical-suspected patients and active retinoblastoma from all treated retinoblastoma patients, the AUC of DLA-RB reached 0.991 (95% CI 0.970-1.000), and 0.962 (95% CI 0.915-1.000), respectively. The combination between ophthalmologists and DLA-RB significantly improved the accuracy of competent ophthalmologists and residents regarding both binary tasks. Cost-utility analysis revealed DLA-RB-based diagnosis mode is cost-effective in both retinoblastoma diagnosis and active retinoblastoma identification. DLA-RB achieved high accuracy and sensitivity in identifying active retinoblastoma from the normal and stable retinoblastoma fundus. It can be used to surveil the activity of retinoblastoma during follow-up and screen high-risk offspring. Compared with referral procedures to ophthalmologic centres, DLA-RB-based screening and surveillance is cost-effective and can be incorporated within telemedicine programs. This study was registered on ClinicalTrials.gov (NCT05308043).

AOP-helpFinder 2.0: Integration of an event-event searches module

To support the use of alternative methods in regulatory assessment of chemical risks, the concept of adverse outcome pathway (AOP) constitutes an important toxicological tool. AOP represents a structured representation of existing knowledge, linking molecular initiating event (MIE) initiated by a prototypical stressor that leads to a cascade of biological key event (KE) to an adverse outcome (AO). Biological information to develop such AOP is very dispersed in various data sources. To increase the chance of capturing relevant existing data to develop a new AOP, the AOP-helpFinder tool was recently implemented to assist researchers to design new AOP. Here, an updated version of AOP-helpFinder proposes novel functionalities. The main one being the implementation of an automatic screening of the abstracts from the PubMed database to identify and extract event-event associations. In addition, a new scoring system was created to classify the identified co-occurred terms (stressor-event or event-event (which represent key event relationships) to help prioritization and support the weight of evidence approach, allowing a global assessment of the strength and reliability of the AOP. Moreover, to facilitate interpretation of the results, visualization options are also proposed. The AOP-helpFinder source code are fully accessible via GitHub, and searches can be performed via a web interface at http://aop-helpfinder-v2.u-paris-sciences.fr/.

Automatic subcutaneous defibrillator screening in patients with intracardiac devices: comparative study depending on the point of pacing

Abstract Funding Acknowledgements Type of funding sources: None. Background/Introduction Subcutaneous defibrillator (S-ICD) has been shown to be effective in the prevention of sudden cardiac death with the advantage of avoiding intravascular access. Although it should not be indicated when pacing indications exists, the need for pacing may arise in the course of underlying heart disease. Before implantation, it is necessary to screen the surface electrocardiogram (ECG) to avoid inappropriate shocks due to overdetection. Published studies so far show a drastic reduction in the percentage of patients eligible for screening with conventional pacing techniques (right ventricular apex, right ventricular outflow tract or resynchronisation with coronary sinus pacing). Purpose The change in screening signal produced by cardiac pacing during is unknown. So far, no studies have been published comparing the outcome of S-ICD screening between different pacing sites, including conduction system pacing (CSP, his and left bundle branch pacing). Our aim is to analyse the success rate of screening and its predictors in patients with cardiac pacing according to the pacing point. Methods A total of 102 patients were screened for S-ICD and divided into 5 groups according to pacing type: His bundle pacing (HBP), left bundle pacing (LBP), right ventricular apex (RVA), right ventricular outflow tract (RVOT) and biventricular cardiac resynchronisation (CRT). There were no significant differences in any variable between the HBP and LBP group, so they were unified as the CSP group. We used a control group of 10 healthy patients with no implantable devices or ECG abnormalities. All patients underwent automatic screening with paced rhythm using the automatic algorithm (Boston Scientific). We considered a suitable screening when at least one vector (primary, secondary or alternative) was positive in both decubitus and standing position, in left or right parasternal position. Results Baseline characteristics by group are summarised in table1. Paced QRS width was lower in the CSP group (135 [120-160] vs 120 [106-120] p=0,001). All healthy controls passed the screening (100%). 39 (97.5%) CSP patients passed the screening, compared to 15 (71.4%) CRT, 14 (70%) RVOT and 4 (19%) RVA. The differences were statistically significant, p < 0.001 (image 1 A & B). In multivariate analysis, only pacing type was shown to be a predictor of positive screening (p 0.001), independent of paced QRS width. Conclusions In our experience, the probability of obtaining a positive screening of the S-ICD is higher with conduction system stimulation than with the other pacing points (AVD, RVOT or CRT).

Multiresolution analysis for COVID-19 diagnosis from chest CT images: wavelet vs. contourlet transforms

Chest computer tomography (CT) provides a readily available and efficient tool for COVID-19 diagnosis. Wavelet and contourlet transforms have the advantages of being localized in both space and time. In addition, multiresolution analysis allows for the separation of relevant image information in the different subbands. In the present study, transform-based features were investigated for COVID-19 classification using chest CT images. Several textural and statistical features were computed from the approximation and detail subbands in order to fully capture disease symptoms in the chest CT images. Initially, multiresolution analysis was performed considering three different wavelet and contourlet levels to determine the transform and decomposition level most suitable for feature extraction. Analysis showed that contourlet features computed from the first decomposition level (L1) led to the most reliable COVID-19 classification results. The complete feature vector was computed in less than 25 ms for a single image having of resolution 256 × 256 pixels. Next, particle swarm optimization (PSO) was implemented to find the best set of L1-Contourlet features for enhanced performance. Accuracy, sensitivity, specificity, precision, and F-score of a 100% were achieved by the reduced feature set using the support vector machine (SVM) classifier. The presented contourlet-based COVID-19 detection method was also shown to outperform several state-of-the-art deep learning approaches from literature. The present study demonstrates the reliability of transform-based features for COVID-19 detection with the advantage of reduced computational complexity. Transform-based features are thus suitable for integration within real-time automatic screening systems used for the initial screening of COVID-19.

Establishment and application of suspension static method in blood group screening of automated blood group analyzer

The accuracy of blood group identification is the basis of blood transfusion safety. In order to increase the detection rate of weak agglutination, unexpected antibodies (UAb) and blood subtypes for pre-transfusion testing, the blood group screening process of automated blood group analyzer (ABGA) is ameliorated by introducing one static step and establishing a suspension static method (SSM). One static step was introduced in the blood group screening process of ABGA and three static time conditions were designed: 300 s, 400 s and 500 s, from which the optimal static time was selected and SSM was established; By comparing the detection of weak agglutination and UAb before and after the application of SSM, the feasibility and effect of suspension static method were verified and evaluated. The last two steps of the automatic blood group screening process were replaced with static, light centrifugation and imaging. The optimal static time parameter was selected as 400 s and SSM was established; After the application of SSM, it was verified that: (1) The detection level of weak antibodies (anti-A and anti-B) and weak antigens (weak D phenotype) could be improved by SSM, including antibodies in plasma of known type O samples with 0, 2, 4, 8, 16 and 32 times serial dilutions(simulating weak anti-A and weak anti-B), weak antibodies (anti-B) in plasma of one normal A-type sample and weak antigens on red blood cells (RBC) of 5 weak D phenotype samples (weak D antigen); (2) Three blood donor samples (type A, O and B) with known UAb were detected by SSM. The results showed that SSM could detect both weak antibodies (anti-A and anti-B) and UAb; (3) SSM was applied to detect the samples of 3 A2B and 3 subtype B blood donors and the blood subtypes could be clearly detected; (4) The number of screening samples was 95,314 and 106,814 before SSM (2018) and after (2020) the application of SSM and the positive rate of UAb (63/95,314 and 187/106,814) increased after SSM, discrepancy of which was statistically significant (χ2 = 48.42, P < 0.01). The above results demonstrate that SSM of ABGA is conducive to the detection of weak agglutination, UAb and blood subtypes in blood samples, which can improve the sensitivity of blood group detection and ensure the safety of clinical blood transfusion to a certain extent.

PO-03-124 DEVELOPMENT AND VALIDATION OF A DEEP NEURAL NETWORK TO MEASURE QTC ON PEDIATRIC ELECTROCARDIOGRAMS

A prolonged Bazett corrected QT (QTcB) either congenital or acquired is one of the primary risk factors in the development of ventricular arrhythmias and sudden cardiac death in children. QTcB determination is notoriously problematic for non-pediatric cardiologists and currently available automatic electrocardiogram (ECG) interpreters. We set out to develop and validate a deep neural network (DNN) to measure QTcB and diagnose prolonged QTc on pediatric ECGs. All resting ECGs from both inpatient and outpatient settings at a quaternary freestanding children’s hospital performed on children 0-18 years of age and read by a pediatric electrophysiologist (PEP) for which the QTcB was manually measured between January 2010 and December 2020 were included. Using the manual QTcB measurements as labels and matrices of all 15 lead voltages as features a convolutional DNN with residuals blocks was trained to measure the QTcB (unweighted DNN) and a dynamic bias was added to optimize the false negative rate and make the model more conservative (dynamically-weighted DNN). The QTcBs on a prospective sample of 200 randomly selected ECGs from 2021 were manually measured by each of 3 PEPs. The ECGs were further classified as pathologically prolonged if the QTcB >=460 ms, and the diagnostic accuracy of the algorithms and the PEPs validated against the PEP consensus estimate of the QTcB. The DNN was trained on a sample of 65,370 ECGs from 37,992 patients. The mean absolute error (MAE) of the PEPs against the consensus of all others was 17.3 (16.2-18.4) ms. Both the unweighted DNN (MAE 10.9 (10.3-11.5) ms) and the dynamically-weighted DNN (MAE 12.8 (12.0-13.6) ms) outperformed the PEPs. The sensitivity of individual PEPs to detect a prolonged QTc (>=460 ms) was 72.7% which was significantly better than the unweighted DNN (64.3%, p=0.045) but significantly worse than the dynamic DNN (92.8%, p<0.001). The positive likelihood ratio of the PEPs was lower than the uneweighted model (p=0.006) but higher than the dynamic model (p=0.002). The negative likelihood ratio of the dynamic DNN (1:12.8) outperformed both the unweighted DNN (1:2.7) and the PEPs (1:3.5). A dynamically-weighted DNN can serve as an automatic screening test to accurately detect pathologically prolonged QTc’s on pediatric ECGs given a higher sensitivity and lower negative likelihood ratio than pediatric electrophysiologists while achieving more accurate estimates of the QTc itself.

Identification of overhead line fault traveling wave and interference clutter based on convolution neural network and random forest fusion

High-speed traveling wave acquisition devices often use a mutation start algorithm with a low threshold value, which can collect a large number of interference clutters. If the devices automatically screen out a flashover fault traveling wave from a massive traveling wave, the traveling wave fault location performance could be improved. In this paper, a fault and interference classification method based on the random forest algorithm, which uses a convolutional neural network as a supervised feature extractor of traveling wave data, is proposed. First, one-dimensional traveling wave data are mapped to a two-dimensional matrix by dividing the information section, and a gray image is used to characterize it. Next, a two-dimensional convolution neural model of traveling wave data is constructed to realize the self-learning of traveling wave data characteristics, and the waveform feature sequence of traveling wave data is obtained. Then, on this basis, a random forest algorithm is used to realize automatic identification and screening of flashover fault traveling waves. Finally, a large number of tests on the simulation and measured data show that the proposed combined algorithm based on the feature extraction and classification of stochastic forest algorithm has better recognition effect of fault and interference than the traditional support vector machine, single stochastic forest algorithm, and convolutional neural network.

Interventions for work participation of unemployed or work-disabled cancer survivors: a systematic review

Background: Supporting unemployed or work-disabled cancer survivors in their work participation can have extensive individual and societal benefits. We aimed to identify and summarise interventions for work participation of unemployed or work-disabled cancer survivors. Methods: Five databases (Medline, Embase, PsycINFO, CINAHL and Cochrane Library) were systematically searched for quantitative studies on interventions aimed at enhancing work participation of unemployed or work-disabled cancer survivors. Work participation refers to participation in the workforce, fulfilling one’s work role. Manual and automatic screening (with ASReview software) were performed on titles and abstracts, followed by manual full-text screening. Data were extracted regarding study, patient and intervention characteristics, and work participation outcomes. Risk of bias (RoB) was assessed using the Cochrane RoB2 and QUIPS tools. Results: We identified 10,771 articles, of which we included two randomised controlled trials (RCTs), of which one feasibility RCT, and three cohort studies. In total, 1862 cancer survivors were included, with predominantly breast cancer. Work participation was mainly measured as time to return to work (RTW) and RTW rate. Interventions included components of coaching (e.g., psychological or rehabilitation), training (e.g., building confidence and managing fatigue) and self-management. Two RCTs with unclear RoB did not show an effect of multicomponent interventions compared to care as usual. One cohort study found a significant effect of a psycho-educational intervention on RTW rates, with moderate RoB. The other two cohort studies, with moderate RoB, reported significant associations between components including job search and placement assistance, and work participation. Discussion: Only few interventions aimed at enhancing work participation of unemployed or work-disabled cancer survivors, have been evaluated. In two cohort studies, promising components for future multicomponent interventions were identified. However, findings suggest that more evidence is necessary on such multicomponent interventions, in which elements explicitly directed at work and including the workplace should be included.

Deep learning-based EEG analysis to classify normal, mild cognitive impairment, and dementia: Algorithms and dataset

For automatic EEG diagnosis, this paper presents a new EEG data set with well-organized clinical annotations called Chung-Ang University Hospital EEG (CAUEEG), which has event history, patient’s age, and corresponding diagnosis labels. We also designed two reliable evaluation tasks for the low-cost, non-invasive diagnosis to detect brain disorders: i) CAUEEG-Dementia with normal, mci, and dementia diagnostic labels and ii) CAUEEG-Abnormal with normal and abnormal. Based on the CAUEEG dataset, this paper proposes a new fully end-to-end deep learning model, called the CAUEEG End-to-end Deep neural Network (CEEDNet). CEEDNet pursues to bring all the functional elements for the EEG analysis in a seamless learnable fashion while restraining non-essential human intervention. Extensive experiments showed that our CEEDNet significantly improves the accuracy compared with existing methods, such as machine learning methods and Ieracitano-CNN (Ieracitano et al., 2019), due to taking full advantage of end-to-end learning. The high ROC-AUC scores of 0.9 on CAUEEG-Dementia and 0.86 on CAUEEG-Abnormal recorded by our CEEDNet models demonstrate that our method can lead potential patients to early diagnosis through automatic screening.

Deep efficient-nets with transfer learning assisted detection of COVID-19 using chest X-ray radiology imaging.

Corona Virus (COVID-19) could be considered as one of the most devastating pandemics of the twenty-first century. The effective and the rapid screening of infected patients could reduce the mortality and even the contagion rate. Chest X-ray radiology could be designed as one of the effective screening techniques for COVID-19 exploration. In this paper, we propose an advanced approach based on deep learning architecture to automatic and effective screening techniques dedicated to the COVID-19 exploration through chest X-ray (CXR) imaging. Despite the success of state-of-the-art deep learning-based models for COVID-19 detection, they might suffer from several problems such as the huge memory and the computational requirement, the overfitting effect, and the high variance. To alleviate these issues, we investigate the Transfer Learning to the Efficient-Nets models. Next, we fine-tuned the whole network to select the optimal hyperparameters. Furthermore, in the preprocessing step, we consider an intensity-normalization method succeeded by some data augmentation techniques to solve the imbalanced dataset classes' issues. The proposed approach has presented a good performance in detecting patients attained by COVID-19 achieving an accuracy rate of 99.0% and 98% respectively using training and testing datasets. A comparative study over a publicly available dataset with the recently published deep-learning-based architectures could attest the proposed approach's performance.

Multistage Optimization of Occupational Safety for Cotton Fabric Screen Printing Press Operators

ABSTRACT The aim of this article is to conduct an analysis leading to the optimization of industrial safety related to the use of a cotton fabric automatic screen printing press. The author’s multistage algorithm, i.e., the analysis-synthesis-minimization procedure, brings effective solutions to the issue in question. There are several risk factors at the printing press workstation and its environment, which are identified by employee surveys, scorecards and airborne samples testing. The risks are visualized by the fishbone/Ishikawa diagram. In accordance with the Fine–Kinney method, a set of 70 scorecards assessing occupational risk is applied to support the identification of critical hazards, which is next re-analyzed using the Analytic Hierarchy Process (AHP). Hierarchically assessed risk factors related to the use of a printing press are easy to be minimized by organizational and technical solutions which are intended to be implemented at the workplace.

Joint optic disc and cup segmentation based on elliptical-like morphological feature and spatial geometry constraint

Glaucoma is a chronic degenerative disease that is the second leading cause of irreversible blindness worldwide. For a precise and automatic screening of glaucoma, detecting the optic disc and cup precisely is significant. In this paper, combining the elliptical-like morphological features of the disc and cup, we reformulate the segmentation task from a perspective of ellipse detection to explicitly segment and directly get the glaucoma screening indicator. We detect the minimum bounding boxes of ellipses firstly, and then learn the ellipse parameters of these regions to achieve optic disc and cup segmentation. Considering the spatial geometry prior knowledge that the cup should be within the disc region, Paired-Box RPN is introduced to simultaneously detect the disc and cup coupled. In addition, boundary attention module is introduced to use edges of the disc and cup as an important guide for context aggregation to improve the accuracy. Comprehensive experiments clearly show that our method outperforms the state-of-the-art methods for optic disc and cup segmentation. Simultaneously, the proposed method also obtains the good glaucoma screening performance with calculated vCDR value. Joint optic disc and cup segmentation, which utilizes the elliptical-like morphological features and spatial geometry constraint, could improve the performance of optic disc and cup segmentation.