• Journal of the Korean Society of Radiology
• /
• v.14 no.1
• /
• pp.39-44
• /
• 2020

The purpose of this study was learning about chest X-ray image classification and accuracy research through Deep Learning using big data technology with Convolution Neural Network. Normal 1,583 and Pneumonia 4,289 were used in chest X-ray images. The data were classified as train (88.8%), validation (0.2%) and test (11%). Constructed as Convolution Layer, Max pooling layer size 2×2, Flatten layer, and Image Data Generator. The number of filters, filter size, drop out, epoch, batch size, and loss function values were set when the Convolution layer were 3 and 4 respectively. The test data verification results showed that the predicted accuracy was 94.67% when the number of filters was 64-128-128-128, filter size 3×3, drop out 0.25, epoch 5, batch size 15, and loss function RMSprop was 4. In this study, the classification of chest X-ray Normal and Pneumonia was predictable with high accuracy, and it is believed to be of great help not only to chest X-ray images but also to other medical images.

• KSBB Journal
• /
• v.16 no.1
• /
• pp.30-35
• /
• 2001

High cell density culture of Bacillus thuringiensis was conducted in fed-batch culture and TCRC using a bioreactor incorporating ceramic membrane filter. Cell growth of B. thuringiensis in fed-batch culture increased linearly, which was well matched by the results of cell growth modeling. In spite of the slower growth rate during fed-batch culture, no spore formation was observed, which was contrary to the results of continuous culture. Changing culture mode to batch culture after fed-batch operation induced a 2.7$\times$$10^9$ CFU/mL spore concentration using a 300 g/L glucose feed concentration. In TCRC operation incorporating ceramic filter within the bioreactor, the effect of glucose feed concentrations on the cell growth and spore formation of B. thuringiensis was determined. A maximum cell concentration of 1.8$\times$$10^{10}$ CFU/ml, which corresponds to 82.6 g-cell/L, was obtained in the TCRC using a 50 g/L glucose feed concentration. In the TCRC, cell growth increased linearly and glucose concentration was limited, which agreed well with the results of cell growth modeling. No spore formation was observed except when 1 g/L of glucose was fed. Changing to batch culture induced a 1.2$\times$$10^{10}$ CFU/mL of spore concentration, which was the highest spore concentration obtained among the various culture modes examined. The optimal glucose feed rate was found to be 0.55 g-glucose/h.

• Journal of Environmental Science International
• /
• v.18 no.6
• /
• pp.691-698
• /
• 2009

Trickling filter has been extensively studied for the domestic wastewater treatment especially for the small scale plants in rural area. The performance of the trickling filter depends on the microbial community and their activity in the biofilms on the media. Nitrification. denitrification, and phosphorus removal of the trickling filter from the wastewater depend on the activity and the amount of the specific microorganisms responsible for the metabolism. For the estimation of the performance of a trickling filter, batch nitrification experiment and fluorescence in situ hybridization (FISH) were carried out to measure the microbial activity and its distribution on the media of the trickling filter. Batch nitrification activity measurement showed that the top part of the 1st stage trickling filter had the highest nitrification activity and the maximum activity was 0.002 g $NH_4$-N/g MLVSS${\cdot}$h. It is thought that higher substrate (ammonia) concentration yields more nitrifying bacteria in the biofilms. The dominant ammonia oxidizer and nitrite oxidizer in the biofilm were Nitrosomonas species and genus Nitrospira, respectively, by FISH analysis. Less denitrifiers were found than nitrifiers in the biofilm by the probe Rrp1088 which specifically binds to Rhodobacter, Rhodovulum, Roseobacter, and Paracoccus. Phosphorus accumulating bacteria were mostly found at the surface of the biofilm by probe Rc988 and PAO651 which specifically binds to Rhodocyclus group and their biomass was less than that of nitrifiers.

• Journal of Astronomy and Space Sciences
• /
• v.34 no.1
• /
• pp.19-30
• /
• 2017

The optical wide-field patrol network (OWL-Net) is a Korean optical surveillance system that tracks and monitors domestic satellites. In this study, a batch least squares algorithm was developed for optical measurements and verified by Monte Carlo simulation and covariance analysis. Potential error sources of OWL-Net, such as noise, bias, and clock errors, were analyzed. There is a linear relation between the estimation accuracy and the noise level, and the accuracy significantly depends on the declination bias. In addition, the time-tagging error significantly degrades the observation accuracy, while the time-synchronization offset corresponds to the orbital motion. The Cartesian state vector and measurement bias were determined using the OWL-Net tracking data of the KOMPSAT-1 and Cryosat-2 satellites. The comparison with known orbital information based on two-line elements (TLE) and the consolidated prediction format (CPF) shows that the orbit determination accuracy is similar to that of TLE. Furthermore, the precision and accuracy of OWL-Net observation data were determined to be tens of arcsec and sub-degree level, respectively.

• International Journal of Aeronautical and Space Sciences
• /
• v.18 no.2
• /
• pp.327-333
• /
• 2017

This study presents the orbit determination (OD) of a candidate Korea Regional Navigation Satellite System (KRNSS) using both inter-satellite links (ISLs) and ground observations. The candidate constellation of KRNSS is first introduced. The OD algorithm based on both ISL and ground observation is developed, and consists of three main components: dynamic model for Korean navigation satellites, measurement model for ISLs and ground observations, and the batch least-square filter for estimating OD parameters. As numerical simulations are performed to analyze the OD performances, the present study focuses on investigating the effects of ISL measurements on the OD accuracy of KRNSS. Simulation results show that the use of ISLs can considerably enhance the OD accuracy to one meter (design preference) under certain distributions of ground stations.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.40 no.2
• /
• pp.108-117
• /
• 2012

This paper focuses on a performance analysis of TRN among various nonlinear filtering methods. In a TRN research, extended Kalman filter(EKF) is a basic estimation algorithm. In this paper, iterated EKF(IEKF), EKF with stochastic linearization(SL), and unscented Kalman filter(UKF) algorithms are introduced to compare navigation performance with original EKF. In addition to introduced sequential filters, bank of Kalman filters method, which is one of the batch method, is also presented. Finally, by simulating an artificial aircraft mission, EKF with SL was chosen as the most consistent filter in the introduced sequential filters. Also, results suggested that the bank of Kalman filters can be alternative for TRN, when a fast convergence of navigation solution is needed.

• ETRI Journal
• /
• v.38 no.6
• /
• pp.1190-1196
• /
• 2016

We propose a new bandpass filter (BPF)-based online channel normalization method to dynamically suppress channel distortion when the speech and channel noise components are unknown. In this method, an adaptive modulation frequency filter is used to perform channel normalization, whereas conventional modulation filtering methods apply the same filter form to each utterance. In this paper, we only normalize the two mel frequency cepstral coefficients (C0 and C1) with large dynamic ranges; the computational complexity is thus decreased, and channel normalization accuracy is improved. Additionally, to update the filter weights dynamically, we normalize the learning rates using the dimensional power of each frame. Our speech recognition experiments using the proposed BPF-based blind channel normalization method show that this approach effectively removes channel distortion and results in only a minor decline in accuracy when online channel normalization processing is used instead of batch processing

• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10a
• /
• pp.299-302
• /
• 1996

A control algorithm is proposed for nonlinear multi-input multi-output(MIMO) batch processes by combining quadratic iterative learning control(Q-ILC) with model predictive control(MPC). Both controls are designed based on output feedback and Kalman filter is incorporated for state estimation. Novelty of the proposed algorithm lies in the facts that, unlike feedback-only control, unknown sustained disturbances which are repeated over batches can be completely rejected and asymptotically perfect tracking is possible for zero random disturbance case even with uncertain process model.

• Journal of Microbiology and Biotechnology
• /
• v.11 no.2
• /
• pp.229-233
• /
• 2001

Trichoderma reesei Rut C-30 produced high levels of ${\beta}$-glucosidase, endo-${\beta}$-glucosidase, endo-${\beta}$-1,4-glucanase, and exo-${\beta}$-1,4-glucanase. In pilot-scale production (50-1 fermentor), productivity and yield of CMCase (carborymethyl cellulose) and FPase (filter paper activity) were 273 U/ml and 35 U/ml, and 162 FPU/l.h and 437 FPU/g, respectively. The fed-batch techniques were used to improve enzyme activities with constant cell concentration. The acidity was an important parameter and controlled at pH 3.9 and 5.0 by automatic addition of ammonium hydroxide. Cellulase powder was prepared by ammonium sulfate precipitation and its CMCase and FPase activities were 3,631 U/g and 407 U/g, respectively.

• Journal of Microbiology and Biotechnology
• /
• v.9 no.1
• /
• pp.44-49
• /
• 1999

Cellulase production by fed-batch cultivation of Trichoderma reesei Rut C30 with various initial concentrations of Solka Floc in 1 % wheat bran-containing medium was investigated. The cellulase activity and productivity increased with initial Solka Floc concentration up to 5%. When a total Solka Floc concentration of 90 g/l was used for cellulase production, CMC (carboxymethyl cellulose) and FP (filter paper) activities, productivity, and yield were 359.7 U/ml, 30.61 U/ml, 161 FPU $L^{-1}$ $h^{-1}$, and 340 FPU $g^{-1}$, respectively. It was important to maintain a high cell concentration during cellulase production to obtain high cellulase activity and productivity. Cellulase powder was prepared by ammonium sulfate precipitation: FP activity was 396.7 U/g and CMC activity was 6481 U/g.