A Deep Neural Network for Detecting Coronavirus Disease Using Chest X-Ray Images

The novel Covid-19 is one of the leading cause of death worldwide in the year 2020 and declared as a pandemic by world health organization (WHO). This virus affecting all countries across the world and 5 lakh people die as of June 2020 due to Covid-19. Due to the highly contagious nature, early detection of this virus plays a vital role to break Covid chain. Recent studies done by China says that chest CT and X-Ray image may be used as a preliminary test for Covid detection. Deep learning-based CNN model can use to detect Coronavirus automatically from the chest X-rays images. This paper proposed a transfer learning-based approach to detect Covid disease. Due to the less number of Covid chest images, we are using a pre-trained model to classify X-ray images into Covid and Normal class. This paper presents the comparative study of a various pre-trained model like VGGNet-19, ResNet50 and Inception_ResNet_V2. Experiment results show that Inception_ResNet_V2 gives the better result as compare to VGGNet and ResNet model with training and test accuracy of 99.26 and 94, respectively.

PENGARUH MODEL PEMBELAJARAN SNOWBALL THROWING TERHADAP PENINGKATAN HASIL BELAJAR SISWA DI SMK NEGERI 3 TONDANO

PENGARUH MODEL PEMBELAJARAN SNOWBALL THROWING TERHADAP PENINGKATAN HASIL BELAJAR SISWA DI SMK NEGERI 3 TONDANO Oleh : David G. Sangari 14212081 Dosen Pembimbing : Drs. H. J. R. Sumarauw, M.Pd Dra. Lenie Ratag, M.Pd ABSTRAK David G. Sangari. 2019. The Influence of the Snowball Throwing Learning Model on the Improvement of Student Learning Outcomes at SMK Negeri 3 Tondano. Thesis, PTM S-1 Study Program, Faculty of Engineering. University Manado State. Advisors (1) Drs. H. J. R. Sumarauw, M.Pd, (II) Dra.Lenie Ratag, M, Pd Keywords: Snowball Throwing Model, Competency Learning of Light Vehicle Engineering Skills Based on the observations of researchers in the field, students in the learning process centered only on teachers and students only memorized concepts. The results at this reflection stage would be the basis for making improvements in this study. This study used 2 classes, namely the experimental class and the control class. The research was conducted in 12 meetings in each class. At the first meeting, a preliminary test was given which aims to see to what extent the students' ability limits. The second meeting until the 11th meeting was given a learning process using the Snowball Throwing learning model in the experimental class and conventional learning models in the control class. In the end, at the 12th meeting, each class was given a posttest to see if there was an improvement from the first meeting. Based on the research hypothesis testing at SMK N 3 Tondano on basic automotive engineering subjects with jacking and blocking material, it can be seen that Snowball Throwing learning affects students' mastery of concepts.

An Investigation on a Plane-Based Dynamic Calibration Method for the Handheld LiDAR Scanner

A plane-based dynamic calibration method had been proposed by the previous study for the GeoSLAM ZEB Horizon handheld LiDAR scanner. Only one preliminary test was presented. Therefore, three datasets in a calibration field were collected in this study on different dates and at different times on the same date to investigate the efficiency of the proposed calibration approach and calibration results. The calibration results for these three datasets showed that all average residuals were closer to 0, and all a posterior unit weight standard deviations of the adjustment were also significantly reduced after calibration. Moreover, the RMSE (root mean square error) of the check planes was improved by about an average of 32.61%, 28.44%, and 14.7%, respectively, for the three datasets. The improvement was highly correlated with the quality of the calibration data. The RMSE differences of all check planes using calibration data collected on different dates and at different times on the same date for calibration was about 1–2 cm and less than 1 mm, respectively. There was no difference in the calibration results, demonstrating the efficiency of the proposed calibration approach and the calibration results during the two different dates.

A Computerized Simulation of the Occlusal Surface in Equine Cheek Teeth: A Simplified Model

Equine mastication, as well as dental wear patterns, is highly important for the development of treatments in equine dentistry. During the last decades, the stress and strain distributions of equine teeth have been successfully simulated using finite element analysis. Yet, to date, there is no simulation available for dental tooth wear in equines. In this study, we developed a simplified two-dimensional computer simulation of dental wear. It provides a first tentative explanation for the development of the marked physiological inclination of the occlusal surface and for pathological conditions such as sharp enamel points in equine cheek teeth. The mechanical properties of the dental structures as well as the movement of the mandible during the equine chewing cycle were simulated according to previously published data. The simulation setup was optimized in preliminary test runs. Further simulations were conducted varying the lateral excursion of the mandible and the presence or absence of incisor contact during the chewing cycle. The results of simulations showed clear analogies to tooth wear patterns in living equids, including the formation of wear abnormalities. Our analysis indicates that small variations in the pattern of movement during the masticatory cycle, as well as incisor contacts, are leading to marked changes in the occlusal tooth wear patterns. This opens new research avenues to better understand the development of dental wear abnormalities in equines and might have serious implications on captive animal health, welfare, and longevity.

Real-time Closed-loop Active Surface Technology of a Large Radio Telescope

The surface accuracy of a large radio telescope’s primary reflector is easily affected by gravity and temperature change during observations. An active surface system is crucial to ensure the regular operation and high-quality data output of the radio telescope. We propose a real-time closed-loop active surface system including two components. The first component, a new type of photoelectric edge sensor, detects the angle change of the adjacent panels. The second component, the displacement actuator, adjusts the panels’ position and posture to compensate for the angle changes. So, over the entire observation, the closed-loop surface control system with these two components could actively maintain the primary reflector’s accuracy in real time. Using this approach, we constructed an experimental active surface system for the Xinjiang Qitai 110 m Radio Telescope (QTT) to test the maintenance of the surface accuracy. The angle measurement accuracy is better than 0.″2, and the positioning accuracy of the displacement actuator could achieve ±15 μm over the whole 50 mm stroke. The preliminary test results show that the accuracy requirements of the QTT’s primary reflector surface can be met using the active surface system we propose.

User-Centric Adaptive Password Policies to Combat Password Fatigue

Today, online users will have an average of 25 password-protected accounts online, yet use, on average, 6.5 passwords. The excessive cognitive burden of remembering large amounts of passwords causes Password Fatigue. Therefore users tend to reuse passwords or recycle password patterns whenever prompted to change their passwords regularly. Researchers have created Adaptive Password Policies to prevent users from creating new passwords similar to previously created ones. However, this approach creates user frustration as it neglects users’ cognitive burden. This paper proposes a novel User-Centric Adaptive Password Policy (UCAPP) Framework for password creation and management that assigns users system-generated passwords based on a cognitive-behavioural agent-based model. The framework comprises a Password Policy Assignment Test (PassPAST), a Cognitive Burden Scale (CBS), a User Profiling Algorithm, and a Password Generator (PassGEN). The framework creates tailor-made password policies that maintain password memorability for users of different cognitive thresholds without sacrificing password strength and entropy. The framework successfully created 30-40% stronger passwords for Critical users and random (non-mnemonic) passwords for Typical users based on each individual’s cognitive password thresholds in a preliminary test.

Acute Toxicity of NH4Cl to Tambra Fish (Tor Tambroides) Fry

Ammonia is one of the nutrient wastes products which can be toxic to all vertebrates, including fish. Usually, ammonia toxicity is expressed as Total Ammonia Nitrogen (TAN, mgN/L). Tambra fish (Tor tambroides) is freshwater fish with important economic value in some areas in Indonesia. According to overfishing and forest degradation, the fish has become threatened. This research aimed to calculate the LC50 value of ammonia to Tambra fish (Tor tambroides) and estimate NOEC, LOEC, and MATC values of ammonia to Tambra fish (Tor tambroides). This study was conducted using the static test method with two steps. The first stage is a preliminary test, where the preliminary test is observed for 24 hours With five concentrations excluding control. Whereas in the second stage that is carried out, observation for 96 hours with four concentrations excluding control. The Probit test is Performed to obtain LC50 value results, then proceed with ammonia analysis using the Salicylate Test Kit method. The results showed that ammonia had acute toxicity with LC50-96 hour value of NH4Cl against Tambra fish fry (Tor tambroides) in the amount of 354.615 mg/L. Ammonia caused acute toxicity to Tambra fish fry (Tor tambroides).The LC50-96 hour value was 354.615 mg/L in the form of Total Ammonia Nitrogen (TAN) or it was equivalent to un-ionized Ammonia (NH3) of 9.8937 mg/L, while values of the Lowest Observed Effect Concentration (LOEC) and No Observed Effect Concentration (NOEC) of ammonia on Tambra fish (Tor tambroides) were 578.24 mg/L and 280.18 mg/L, respectively.

Implementation of Distance Learning Based on Problem-Based Learning to Increase Critical Thinking Skills

Problem-based learning (PBL) is a learning model that requires the active participation of each individual, who in the learning process all help to determine the subject matter. The PBL model can foster students’ critical thinking skills through solving various real-world problems found in the environment where they live. This research aimed to improve students’ critical thinking skills through PBL-based online learning. Quantitative experimental methods were used. A one group pretest-posttest design was employed: a preliminary test (pretest) was carried out, followed by the treatment, and finally the final test (posttest). There were two experimental groups: A and B. The results showed that students’ critical thinking skills increased in both group A and group B and this change was significant, as measured through a t-test (p < 0.05 in both groups). Keywords: distance learning; problem-based learning; critical thinking skills; sociology of education.