Exhaustive Process Research Articles

Comparative Analysis of Vanilla CNN and Transfer Learning Models for Glaucoma Detection

Glaucoma is a leading cause of blindness worldwide and results from high eye pressure‐induced damage to the optic nerves, thereby preventing visual information from reaching the brain. While glaucoma is incurable in its advanced stages, its early detection improves the treatment outcome. Recently, computational models for image classification have enabled early detection of glaucoma from OCT scans of patients and will potentially augment the medical diagnosis of this disease. Models that are based on the convolutional neural networks (CNNs) have shown promise in the early detection of glaucoma. These models vary in their architectures and their accuracies depend largely on the intrinsic nature of the training datasets used. Hence, in this work, a comparative analysis is performed on vanilla CNN, AlexNet, GoogLeNet, and ResNet50 using two popular glaucoma datasets (ACRIMA and ORIGA). With careful attention to exhaustive image processing, an impressive training accuracy of 89% and validation accuracy of 50% was obtained from the vanilla CNN, showing a high sensitivity of 88% in detecting glaucomatous patients from OCT scans. However, the other models (AlexNet, GoogLeNet, and ResNet5) overfitted with a large difference between the obtained training accuracy and validation accuracy. The results also reveal that ResNet50 has the highest computational cost compared to the rest of the models. The obtained results demonstrate the peculiarity of the dataset, its selectiveness of the most appropriate model, and the potential of deep neural networks (DNNs) as an effective screening tool for glaucoma, enabling prompt interventions, reducing healthcare costs, and helping optometrists make swift decisions.

Convex Hull Prediction for Adaptive Video Streaming by Recurrent Learning.

Adaptive video streaming relies on the construction of efficient bitrate ladders to deliver the best possible visual quality to viewers under bandwidth constraints. The traditional method of content dependent bitrate ladder selection requires a video shot to be pre-encoded with multiple encoding parameters to find the optimal operating points given by the convex hull of the resulting rate-quality curves. However, this pre-encoding step is equivalent to an exhaustive search process over the space of possible encoding parameters, which causes significant overhead in terms of both computation and time expenditure. To reduce this overhead, we propose a deep learning based method of content aware convex hull prediction. We employ a recurrent convolutional network (RCN) to implicitly analyze the spatiotemporal complexity of video shots in order to predict their convex hulls. A two-step transfer learning scheme is adopted to train our proposed RCN-Hull model, which ensures sufficient content diversity to analyze scene complexity, while also making it possible to capture the scene statistics of pristine source videos. Our experimental results reveal that our proposed model yields better approximations of the optimal convex hulls, and offers competitive time savings as compared to existing approaches. On average, the pre-encoding time was reduced by 53.8% by our method, while the average Bjøntegaard delta bitrate (BD-rate) of the predicted convex hulls against ground truth was 0.26%, and the mean absolute deviation of the BD-rate distribution was 0.57%.

Особенности нестационарных тепломеханических процессов в выпускном коллекторе автомобильного поршневого двигателя размерности 8,2/7,1

The purpose of this work was to obtain experimental data on unsteady gas dynamics and the intensity of heat transfer of the flow in the exhaust system of an automobile piston engine for different initial conditions. The research was carried out on an experimental stand simulating the exhaust process in a piston engine in dynamics. The pressure at the outlet (in the cylinder) varied from 0.05 to 0.2 MPa, the speed of rotation of the crankshaft - from 600 to 3000 min-1. The method of thermal anemometry was used to determine the gas dynamic, flow and heat exchange characteristics of the flow in the exhaust system. Data on the instantaneous values of the velocity and local heat transfer of the flow in the exhaust system for different boundary conditions are presented in the article. The effect of the outlet pressure on the gas flow through the exhaust system and the level of heat transfer was analyzed. The spectral analysis of changes in the local flow velocity in the exhaust system is additionally presented in the article. The obtained patterns are useful in the applied aspect for clarifying engineering methods for calculating the exhaust system and for verifying mathematical models of gas exchange processes in automotive piston engines.

A Novel Gender Classification Model based on Convolutional Neural Network through Handwritten Text and Numeral

Human handwriting is used to investigate human characteristics in various applications, including but not limited to biometric authentication, personality profiling, historical document analysis, and forensic investigations. Gender is one of the most distinguishing characteristics of human beings. From this point forth, we propose a novel end-to-end model based on Convolutional Neural Network (CNN) that automatically extracts features from a given handwritten sample, which contains both handwritten text and numerals unlike the related work that uses only handwritten text, and classifies its owner’s gender. In addition to proposing a novel model, we introduce a new dataset that consists of 530 gender-labeled Turkish handwritten samples since, to the best of our knowledge, there does not exist a public gender-labeled Turkish handwriting dataset. Following an exhaustive process of hyperparameter optimization, the proposed CNN featured the most optimal hyperparameters and was both trained and evaluated on this dataset. According to the experimental result, the proposed novel model obtained an accuracy as high as 74.46%, which overperformed the state-of-the-art baselines and is promising on such a task that even humans could not have achieved highly-accurate results for, as of yet.

A multistage Sendai virus vaccine incorporating latency-associated antigens induces protection against acute and latent tuberculosis

ABSTRACT One-quarter of the world’s population is infected with Mycobacterium tuberculosis (Mtb). After initial exposure, more immune-competent persons develop asymptomatic latent tuberculosis infection (LTBI) but not active diseases, creates an extensive reservoir at risk of developing active tuberculosis. Previously, we constructed a novel recombinant Sendai virus (SeV)-vectored vaccine encoding two dominant antigens of Mtb, which elicited immune protection against acute Mtb infection. In this study, nine Mtb latency-associated antigens were screened as potential supplementary vaccine candidate antigens, and three antigens (Rv2029c, Rv2028c, and Rv3126c) were selected based on their immune-therapeutic effect in mice, and their elevated immune responses in LTBI human populations. Then, a recombinant SeV-vectored vaccine, termed SeV986A, that expresses three latency-associated antigens and Ag85A was constructed. In murine models, the doses, titers, and inoculation sites of SeV986A were optimized, and its immunogenicity in BCG-primed and BCG-naive mice were determined. Enhanced immune protection against the Mtb challenge was shown in both acute-infection and latent-infection murine models. The expression levels of several T-cell exhaustion markers were significantly lower in the SeV986A-vaccinated group, suggesting that the expression of latency-associated antigens inhibited the T-cell exhaustion process in LTBI infection. Hence, the multistage quarter-antigenic SeV986A vaccine holds considerable promise as a novel post-exposure prophylaxis vaccine against tuberculosis.

Qualitative and quantitative chemometric modelling of nanostructured carbon samples based on infrared spectroscopy

Rapid population growth necessitates a continuous increase in industrial productivity, with a concomitant environmental burden. During the past few years, nanostructured carbon materials have proved their effectiveness in reducing the usage of several hazardous substances, owing to their distinct characteristics. These properties depend on the type of feedstock and the parameters of pyrolysis. We have developed multivariate prediction models to determine several physical properties of nanostructured carbon samples, which are usually calculated from the adsorption isotherms. Adsorption measurement is a time-consuming and exhaustive process. Therefore, our goal was to provide a fast and environmentally friendly alternative to current methods to determine micropore volume, specific surface area and total pore volume, based on FTIR spectroscopy coupled with chemometric methods. Moreover, we have created classification models, which are capable of predicting the used feedstock materials from IR spectra. Our support vector machine–based classification model had the best accuracy values, above 0.86. Our classification and regression models have excellent performance and were properly validated, thus they are good alternatives for the robust and fast determination of the important qualitative and quantitative features of carbon samples.

Investigation of gas and plasma driven deuterium permeation through a vanadium membrane for particle exhaust application in a fusion reactor

The process of particle exhaust based on the cryopump in fusion reactor presents a severe issue of deuterium and tritium retention. However, the metal foil permeation pump represents an effective means to solve the problem of deuterium and tritium retention during the particle exhaust process. How to enhance the permeability performance of metal films to meet the requirements of practical engineering applications of fusion reactor is the main content of the research. This article focuses on the simulation and experimental investigation of V membrane, aiming to explore methods for enhancing membrane permeation flux. Through simulation analysis, the influence of thickness, membrane surface state, and temperature on both plasma-driven and gas-driven permeation has been studied. Among these, the downstream surface condition is changed during the PDP process, resulting in an almost twofold increase in permeation flux. Based on the permeation experiments the influence of temperature on permeation was proved and the impact of plasma source power on permeation was also studied. It was found that the permeation flux increased more than two orders of magnitude when the radio frequency power changed from 0 to 500 W, which indicates that plasma incident flux serves as a more effective means to enhance permeation flux compared to temperature.

Research and Optimization on the Exhaust Flow Characteristics Based on Energy-Splitting Method of the Low-Speed Marine Diesel Engine

Abstract In this study, an optimization scheme for exhaust flow characteristics based on the energy-splitting method is proposed. The low-speed marine diesel engine adopts the exhaust energy-splitting method during the exhaust process. Based on this, a one-dimensional calculation model, a three-dimensional calculation model, and a three-dimensional model of the gas collection box are established. After splitting, the vortex structure in the same phase of the gas collection box has a larger scale, higher flowrate, more significant entropy increase, and more severe turbulence dissipation. The length and diameter of the high-grade gas collection box are optimized. The results show that after optimization, the flow energy dissipation brought by the vortex structure is reduced, and the outlet pressure and flow velocity of the gas collection box are increased.

A Secret Key Classification Framework of Symmetric Encryption Algorithm Based on Deep Transfer Learning

The leakage signals, including electromagnetic, energy, time, and temperature, generated during the operation of password devices contain highly correlated key information, which leads to security vulnerabilities. In traditional encryption algorithms, the length of the key greatly affects the upper limit of its security against cracking. Regarding side-channel attacks on long-key algorithms, traditional template attack methods characterize the energy traces using multivariate Gaussian distribution during the template construction phase. The exhaustive key-guessing process is expected to consume a significant amount of time and computational resources. Therefore, to analyze the effectiveness of obtaining key values from the side information of password devices, we propose an innovative attack method based on a divide-and-conquer logical structure, targeting semi-bytes. We construct a collection of key classification submodules with symmetric correlations. By integrating a differential network model for byte-block sets and an end-to-end direct attack method, we form a holistic symmetric decision framework and propose a key classification structure based on deep transfer learning. This structure consists of three main parts: side information data acquisition, analysis of key-value effectiveness, and determination of attack positions. It employs multiple parallel symmetric subnetworks, effectively improving attack efficiency and reducing the key enumeration range. Experimental results show that the optimal attack accuracy of the network model can reach 91%, with an average attack accuracy of 78%. It overcomes overfitting issues under small sample dataset conditions.

Divided Exhaust Period Assessment for Fuel-Enrichment Reduction in Turbocharged Spark-Ignition Engines

<div>Turbocharged spark-ignition (SI) engines, owing to frequent engine knocking events, utilize retarded spark timing that causes combustion inefficiency, and high turbine inlet temperature (Trb-In T) levels. Fuel enrichment is implemented at high power levels to prevent excessive Trb-In T levels, resulting in an additional fueling penalty and higher CO emissions. In current times, fuel-enrichment reductions are of high strategic importance for engine manufacturers to meet the imminent emissions regulations. To that end, the authors investigated the divided exhaust period (DEP) concept in a 2.2 L turbocharged SI engine with a geometric compression ratio of 14 by decoupling blowdown (BD) and scavenge (SC) events during the exhaust process. Using a validated 1D engine model, the authors first analyzed the DEP concept in terms of pumping mean effective pressure (PMEP) and engine knocking (KI) reduction. Subsequently, the authors examined the effectiveness of the DEP concept using a “low-restriction exhaust flowpath” and varying late intake valve closing (LIVC) duration.</div> <div>First, using DEP, significant PMEP and KI reductions benefits were observed at high power engine conditions along with a large increase in Trb-In T from the early blowdown event. Subsequently, use of a low restriction exhaust flowpath and a shortened LIVC duration further elevated the DEP benefits, including Trb-In T reduction that facilitated enrichment reduction. At 4,000 RPM/20 bar BMEP, ~70% lower PMEP and a 2.2 point increase in ITEg were noted relative to the base engine. However, the 2,000 RPM peak torque engine condition was compromised using DEP, due to knock limitation and deteriorated stock turbocharger performance. Finally, DEP design integrated with an off-the-shelf (new) turbocharger system remedied the low-end torque challenges and demonstrated a notable enrichment reduction and thermal efficiency benefits at the full load engine curve including the 200 kW rated condition.</div>

COVID-19 vaccine and test data management – practice report from a university hospital

Abstract Issue The COVID-19 pandemic generated the urgency for creating and implementing new tools, specifically for data collection, sharing. The data governance of healthcare workers (HCW) was layered on top of pre-existing digitalisation issues in health authorities as well as in health facilities. Description of the problem The vaccination act for the German healthcare sector adopted in December 2021, established the compulsory report of HCWs’ COVID-19 vaccination status. The LMU University Hospital management required an overview of the COVID-19 status (illness, isolation, vaccination, infection) of HCWs in order to ensure continuous clinical service. For the timely and accurate report, it was necessary to create a centralised real-time overview of data that were collected and managed by decentralised structures. Results We connected five in-hospital systems into one platform that was being synced three times a day. For each HCW, we created a dataset with three main categories: core personnel data, COVID-19 status and reason for absence. Two user groups were defined: strategic management (1) and operational management (2), where users of group 2 were only given access to their own unit and were required to update and verify data of their team. Overall, the platform contained data of 14,419 HCWs and had 656 users (Ngroup1 = 16, Ngroup2 = 640). Lessons The data merger revealed issues in interoperability and accuracy and thus expedited data standardisation. The decentralised approach facilitated the operational efficiency, digitalisation, service delivery and managerial empowerment at the hospital. The main challenges pertained to data accuracy and to the continuously changing regulatory requirements as well as privacy protection. As all data were previously hosted on different platforms, an exhaustive process of cleansing was needed prior to merging. Each HCW was assigned to an organisational unit with a respective responsible person in order to achieve data quality. Key messages • Dedicated digital human resources solutions in a university hospital pose a specific challenge for the balance between data privacy, availability and interoperability. • To enable a centralised information display, data must be collected where it originates and made available through short paths. Our solution united decentralised collection and centralised management.

Modeling the Tokamak Exhaust Processing System in a Commercial Simulator for Process Monitoring Purposes

Nuclear fusion depends on tritium breeding and self-sufficiency. Tritium represents a hazard due to its radioactivity and migration properties. Because of these difficulties, ITER, the largest fusion experiment so far, relies on a conservative static procedure to monitor the tritium inventory. Future commercial fusion plants can avoid operation halts if a dynamic monitoring strategy proves itself valid. Tritium plant models have been developed for this kind of monitoring and analysis task, but sensor accuracy and reliability are an issue still to be addressed, and the path to dynamic monitoring remains unclear. The present work shows the modeling procedure of the Tokamak Exhaust Processing system in a commercial simulator, Aspen HYSYS, to reproduce the inventories, streams, process conditions, and compositions of this subsystem during operation. The model is verified in a steady-state scenario using data from the available literature. A demonstration of such a tritium plant subsystem shows meaningful value for several reasons. First, this process has not been modeled before in commercial dynamic simulators, which are typically used in the process industry. It will also allow new stakeholders to participate in future fusion-related projects. Second, it will play a key role in industry-like tritium process monitoring, in which the new model will act as a digital twin of the plant. Data-driven diagnostics can be fueled by model data, helping engineers to generate additional data that could otherwise be expensive to get directly from the plant. For these reasons, models will represent an essential part of a dynamic monitoring system, necessary for feasible fusion projects.

Formability improvement of hot embossing for functional micro-array structures with a novel exhaust punch

Surface microstructures can endow materials with excellent functions such as enhanced heat transfer. However, with the continuous improvement of function enhancement, it is urgent to develop the manufacturing technology of more complex multiscale functional microstructures. Herein, an exhaust hot embossing based on a novel multi-block assembled punch is developed to fabricate micro-array structures, and the microformability of conventional process and exhaust process is compared. The combined processing technology of slow-feeding wire cut electrical discharge machining and micro electrical discharge machining is adopted to manufacturing the assembled exhaust punch with numerous micro-holes and complex multiscale features. The processing quality of the assembled punch is better than the integrated punch, and the surface roughness of each position of the assembled punch is Ra ＜ 0.2 μm. During hot embossing experiments, it is found that the exhaust process could significantly improve the filling performance. At 300 °C, the exhaust process realizes complete filling of the microstructures under the maximum load of 25 kN, whereas the conventional process requires 50 kN. Meanwhile, it is also found that the exhaust process can significantly improve warping and surface coarsening defects compared to the conventional process. The pool boiling test was carried out on the fabricated micro-array structures, and it was found that its heat transfer performance is significantly enhanced compared with other tested surfaces, which will be applied in enhanced heat transfer.

Does the COVID-19 Pandemic Have an Impact on Compassion Fatigue among Midwives?

Professional health services have been key in the recovery of patients during the COVID-19 pandemic. Health professionals have been in a great effort to meet the care needs of patients despite experiencing a high emotional burden for reasons such as unknown treatment of the disease, high contagiousness, changing working systems, increased workload, and the need to wear protective equipment. The unknown effects of the virus on the health of mothers and babies increased the concerns of midwives helping families in the core of society, which are the most intimate areas, regarding the risk of infection. Midwives had to manage an exhausting process similar to the one experienced by nurses, when they continued providing secondary health services such as immunization, antenatal care, and follow-ups. They are among the health professionals forced to work under highly adverse conditions, such as increased duties and responsibilities in the pandemic, and cancellation of leaves. It is reported that midwives, who are forced to work outside their field of duty due to the pandemic, are highly likely to experience compassion fatigue and to have a low level of life quality as well as professional belonging. Qualitative studies conducted with midwives working in the delivery room reported that midwives in delivery rooms experience compassion fatigue, adopt the concept of compassion fatigue, come up with their own solutions or receive psychological support to deal with compassion fatigue, and that they suggested that psychological support needs be provided by the state. As a result, many mothers and pregnant women lost their lives in this ongoing pandemic process. Alike, many health professionals lost their lives due to the epidemic. Midwives are one of these health professionals. The feeling of helplessness at the beginning of the pandemic, the unknowns about its treatment made the working conditions even more difficult. Considering these, it is reasonable to argue that midwives have been likely to experience compassion fatigue.

Chinese holistic processing: Evidence from cognitive mental architecture using Systems Factorial Technology

Previous research has presented conflicting evidence regarding whether Chinese characters are processed holistically. In past work, we applied Systems Factorial Technology (SFT) and discovered that native Chinese speakers exhibited limited capacity when processing characters and words. To pinpoint the source of this limitation, our current research delved further into the mental architecture involved in processing Chinese characters and English words, taking into consideration information from each component. In our current study, participants were directed to make the same/different judgments on characters/words presented sequentially. Our results indicated that participants utilized a parallel self-terminating strategy when both or neither of the left/right components differed (Experiment 1). Faced with the decisional uncertainty that either the left/right component would also differ, most participants processed with a parallel exhaustive architecture, while a few exhibited the coactive architecture (Experiment 2). Taken together, our work provides evidence that in word/character perception, there is weak holistic processing (parallel self-terminating processing) when partial information is sufficient for the decision; robust holistic processing (coactive or parallel exhaustive processing) occurs under decisional uncertainty. Our findings underscore the significant role that the task and presentation context play in visual word processing.

Electron counting detectors in scanning transmission electron microscopy via hardware signal processing

Transmission electron microscopy is a pivotal instrument in materials and biological sciences due to its ability to provide local structural and spectroscopic information on a wide range of materials. However, the electron detectors used in scanning transmission electron microscopy are often unable to provide quantified information, that is the number of electrons impacting the detector, without exhaustive calibration and processing. This results in arbitrary signal values with slow response times that cannot be used for quantification or comparison to simulations. Here we demonstrate and optimise a hardware signal processing approach to augment electron detectors to perform single electron counting.

A Deep Learning Framework for Soft Robots with Synthetic Data.

Data-driven methods with deep neural networks demonstrate promising results for accurate modeling in soft robots. However, deep neural network models rely on voluminous data in discovering the complex and nonlinear representations inherent in soft robots. Consequently, while it is not always possible, a substantial amount of effort is required for data acquisition, labeling, and annotation. This article introduces a data-driven learning framework based on synthetic data to circumvent the exhaustive data collection process. More specifically, we propose a novel time series generative adversarial network with a self-attention mechanism, Transformer TimeGAN (TTGAN) to precisely learn the complex dynamics of a soft robot. On top of that, the TTGAN is incorporated with a conditioning network that enables it to produce synthetic data for specific soft robot behaviors. The proposed framework is verified on a widely used pneumatic-based soft gripper as an exemplary experimental setup. Experimental results demonstrate that the TTGAN generates synthetic time series data with realistic soft robot dynamics. Critically, a combination of the synthetic and only partially available original data produces a data-driven model with estimation accuracy comparable to models obtained from using complete original data.

AUTOMATED ACUTE LYMPHOBLASTIC LEUKEMIA CELL CLASSIFICATION USING OPTIMIZED CONVOLUTIONAL NEURAL NETWORK

Acute lymphoblastic leukemia (ALL) is the most common variant of paediatric cancer that creates numerous immature white blood cells affecting the bone marrow. Manual diagnosis of leukemia from microscopic evaluation of stained sample slides is an exhausting process, which is less accurate and susceptible to human errors. Additionally, identifying the leukemic blast cells under the microscope is complicated due to morphological similarity with the normal cell images. In this paper, we proposed an automated method to analyse the blood smear images using Local Binary Pattern (LBP) and classify the leukemic blast cells and normal cells. We have analysed the performance of machine learning and deep learning models such as Support Vector Machine (SVM), k-Nearest Neighbor algorithm (kNN), Artificial Neural Network (ANN), and Convolutional Neural Network (CNN). For classifying ALL and normal cell images, kNN achieved an accuracy of 94.4%, SVM, and ANN achieved an accuracy of 98.6%, and CNN achieved an accuracy of 99.6%. SVM achieved the highest sensitivity of 100%.

Overcoming Environmental Pollution Due to Waste from the Soun Noodle Processing Industry


 
 
 Pollution is not only caused by large and medium scale industries, but can also be caused by small-scale industries, one of which adalahindustri glass noodles noodles which also produces liquid waste and solid waste. The problem in this research are: How is control environmental pollution caused by the production of cottage industry "soun noodles" in the district of Klaten regency Tulung, and What are the constraints experienced in the process of controlling environmental pollution caused by the production of cottage industry "soun noodles" in the district of Klaten regency Tulung. This research is empirical laws, the data obtained through interviews with respondents and informants, and supported by secondary data. Waste Liquid emanating from the exhaust process chlorine and solid waste in the form of starch fibers discharged into rivers, causing pollution of river water. In this case the maximum BLH not deal with the issue because of the lack of supervision and bugdet. Most of the small-scale glass noodles noodle industry, and are not required to have TDI, so that apart from oversight. In addition, the socialization of the impact of industrial activities on the environment pollution is not maximized, due to limited funding and the number of officers BLH. Based on these results, it recommends that the Office of the investments and unified licensing and BLH should be continuous monitoring and counseling to perpetrators of the business activities and BLH encourage businesses to conduct joint IPAL
 
 

Harnessing generative grammars and genetic algorithms for immersive 2D maps

Modern games typically demand varied, detailed, and wide maps. In particular, 2D game maps are typically designed as a combination of smaller pieces called tiles, so positioning each tile on a map layer is analogous to colorizing a pixel using a palette. However, building a map is an exhaustive process that consumes a lot of planning and maintenance time from a multidisciplinary team, including artists and designers. Consequently, PCG techniques emerge as an attractive solution to this challenge. We propose a novel PCG technique to assist designers in writing rules applied to a balanced combination of genetic algorithms and generative grammar. Our implementation supports generating two-dimensional pathways and places detailed objects (e.g., plants, flowers, forest, etc.) along the path while following a desired distribution given beforehand by the designer. Notably, our PCG method allows the generation of variable transitions of biomes around a road and these results are compatible with real-world technologies for game development and prototyping. Our method can aid designers in various specific modeling tasks regarding environments with two or more dimensions since the results are sufficiently interpretable to undergo subsequent human intervention.