• Journal of the Korean Society of Radiology
• /
• v.18 no.5
• /
• pp.531-540
• /
• 2024

Recently, research and development on automatic diagnosis solutions in the medical imaging field using deep learning are actively underway. In this study, we sought to find a fast and accurate classification deep learning modeling for classification of pneumonia in chest images using Inception V3, a deep learning model based on a convolutional artificial neural network. For this reason, after applying the optimization algorithms AdaGrad, RMS Prop, and Adam to deep learning modeling, deep learning modeling was implemented by selectively applying learning rates of 0.01 and 0.001, and then the performance of chest X-ray image pneumonia classification was compared and evaluated. As a result of the study, in verification modeling that can evaluate the performance of the classification model and the learning state of the artificial neural network, it was found that the performance of deep learning modeling for classification of the presence or absence of pneumonia in chest X-ray images was the best when applying Adam as the optimization algorithm with a learning rate of 0.001. I was able to. And in the case of Adam, which is mainly applied as an optimization algorithm when designing deep learning modeling, it showed excellent performance and excellent metric results when selectively applying learning rates of 0.01 and 0.001. In the metric evaluation of test modeling, AdaGrad, which applied a learning rate of 0.1, showed the best results. Based on these results, when designing deep learning modeling for binary-based medical image classification, in order to expect quick and accurate performance, a learning rate of 0.01 is preferentially applied when applying Adam as an optimization algorithm, and a learning rate of 0.01 is preferentially applied when applying AdaGrad. I recommend doing this. In addition, it is expected that the results of this study will be presented as basic data during similar research in the future, and it is expected to be used as useful data in the health and bio industries for the purpose of automatic diagnosis of medical images using deep learning.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.8
• /
• pp.3529-3550
• /
• 2016

In heterogeneous networks (HetNets), energy saving is vital for a sustainable network development. Many techniques, such as spectrum allocation, network planning, etc., are used to improve the network energy efficiency (EE). In this paper, micro BSs utilizing cell range expansion (CRE) and spectrum allocation are considered in multi-channel heterogeneous networks to improve EE. Hotspot region is assumed to be covered by micro BSs which can ensure that the hotspot capacity is greater than the average demand of hotspot users. The expressions of network energy efficiency are derived under shared, orthogonal and hybrid subchannel allocation schemes, respectively. Particle swarm optimization (PSO) algorithm is used to solve the optimal ratio of subchannel allocation in orthogonal and hybrid schemes. Based on the results of the optimal analysis, we propose three service control strategies on the basis of large-scale user behaviors, i.e., adjust micro cell rang expansion (AmCRE), adjust micro BSs density (AmBD) and adjust micro BSs transmit power (AmBTP). Both theoretical and simulation results show that using shared subchannel allocation scheme in AmBD strategies can obtain maximal EE with a very small area ratio. Using orthogonal subchannel allocation scheme in AmCRE strategies can obtain maximal EE when area ratio is larger. Using hybrid subchannel allocation scheme in AmCRE strategies can obtain maximal EE when area ratio is large enough. No matter which service control strategy is used, orthogonal spectrum scheme can obtain the maximal hotspot user rates.

• Journal of Korean Society of Water and Wastewater
• /
• v.12 no.2
• /
• pp.90-98
• /
• 1998

The optimal rechlorination in water distribution systems was investigated by incorporating optimization techniques into a numerical water quality model. For a hypothetical system that consists of 10 junctions including a storage tank and 12 links, the bulk ($k_b$) and pipe-wall ($k_w$) decay-rate constants of chlorine residual are assumed to be 2.0 1/day and 1.5 m/day, respectively. It was also assumed that the lower and upper limits of chlorine residual in the network are 0.2 mg/L and 0.6 mg/L. When the chlorine source is only the storage tank (without rechlorination), the high levels of chlorine residual appear near the storage tank to maintain the chlorine residuals above the lower limit over the junctions. On the other hand, the chlorine residuals in the network are distribute within the desirable range (0.2 - 0.6 mg/L) after the optimal rechlorination through five injection sites including the storage tank. In case of a real water distribution system that comprises 28 junctions including a clear well and 27 links, the bulk and pipe-wall decay-rate constants are 0.3 1/day and 0.2 m/day, respectively. Before rechlorination, the required chlorine residual at the clearwell is 5.1 mg/L to keep the chlorine residuals above the minimum level (0.6 mg/L) over the junctions. By the optimal rechlorination at five injection sites, the chlorine residuals are distributed within a desirable range of 0.6 mg/L through 2.0 mg/L, which can avoid the excess of chlorine residuals near the clear well. Consequently, total chlirine doses are decreased by 81% in the hypothetical distribution network and 69 % in the real distribution network for satisfying the minimum chlorine residuals.

• Journal of the Korean Society of Hazard Mitigation
• /
• v.10 no.6
• /
• pp.81-87
• /
• 2010

In order to calculate the accurate runoff from a basin, nonlinearity in the relationship between rainfall and runoff has to be considered. Many runoff calculation models assume the linearity in the relationship or are too complicated to be analyzed. Therefore, the storage function method has been used in the prediction of flood because of the simplicity of the model. The storage function method has five parameters with related to the basin and rainfall characteristics which can be estimated by the empirical trial and error method. To optimize these parameters, regression method or optimization techniques such as genetic algorithm have been used, however, it is not easy to optimize them because of the complexity of the method. In this study, the metamodel is proposed to estimate those model parameters. The metamodel is the combination of artificial neural network and genetic algorithm. The model is consisted of two stages. In the first stage, an artificial neural network is constructed using the given rainfall-runoff relationship. In the second stage, the parameters of the storage function method are estimated using genetic algorithm and the trained artificial neural network. The proposed metamodel is applied in the Peong Chang River basin and the results are presented.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.48 no.7
• /
• pp.73-86
• /
• 2006

Concentrated poultry production in the State of Alabama, U.S.A. results in excessive poultry litter. Application of poultry litter to pastures and row crops serves as a cheap alternative to commercial fertilizer. However, over the years, poultry litter application to perennial forage crops in the Appalachian Plateau region of North Alabama has resulted in phosphorus (P) buildup in soils. Phosphorus index (P-index) and comprehensive nutrient management plans (CNMP) are often used as a best management practice (BMP) for proper land application of litter. Because nutrient management planning is often not done for small animal feeding operations (AFOs), and also because, in case of excess litter, litter transportation infrastructure has not been developed, over application of poultry litter to near by area is a common practice. To alleviate this problem, optimal poultry litter management and transportation infrastructure needs to be developed. This paper presents a methodology to optimize poultry litter application and transportation through efficient nutrient management planning and transportation network analysis. The goal was accomplished through implementation of three important modules, a P-Index module, a CNMP module, and a transportation network analysis module within ArcGIS, a Geographic Information System (GIS). The CNMP and P-Index modules assist with land application of poultry litter at a rate that is protective of water quality, while the transportation network analysis module helps transport excess litter to areas requiring litter in the Appalachian Plateau and Black Belt (a nutrient-deficient area) regions. Once fully developed and implemented, such a system will help alleviate water quality problems in the Appalachian Plateau region and poor soil fertility problems in the Black Belt region by optimizing land application and transportation. The utility of the methodology is illustrated through a hypothetical case study.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.50 no.7
• /
• pp.131-139
• /
• 2013

To overcome the limitations in performance and power consumption of traditional electrical interconnection based network-on-chips (NoCs), a hybrid optical network-on-chip (HONoC) architecture using optical interconnects is emerging. However, the HONoC architecture should use circuit-switching scheme owing to the overhead by optical devices, which worsens the latency unfairness problem caused by frequent path collisions. This resultingly exert a bad influence in overall performance of the system. In this paper, we propose a new task mapping algorithm for optimizing latency by reducing path collisions. The proposed algorithm allocates a task to a certain processing element (PE) for the purpose of minimizing path collisions and worst case latencies. Compared to the random mapping technique and the bandwidth-constrained mapping technique, simulation results show the reduction in latency by 43% and 61% in average for each $4{\times}4$ and $8{\times}8$ mesh topology, respectively.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.10
• /
• pp.4738-4753
• /
• 2018

Efficient rendezvous node selection and routing algorithm (RNSRA) for wireless sensor networks with mobile sink that visits rendezvous node to gather data from sensor nodes is proposed. In order to plan an optimal moving tour for mobile sink and avoid energy hole problem, we develop the RNSRA to find optimal rendezvous nodes (RN) for the mobile sink to visit. The RNSRA can select the set of RNs to act as store points for the mobile sink, and search for the optimal multi-hop path between source nodes and rendezvous node, so that the rendezvous node could gather information from sensor nodes periodically. Fitness function with several factors is calculated to find suitable RNs from sensor nodes, and the artificial bee colony optimization algorithm (ABC) is used to optimize the selection of optimal multi-hop path, in order to forward data to the nearest RN. Therefore the energy consumption of sensor nodes is minimized and balanced. Our method is validated by extensive simulations and illustrates the novel capability for maintaining the network robustness against sink moving problem, the results show that the RNSRA could reduce energy consumption by 6% and increase network lifetime by 5% as comparing with several existing algorithms.

• Journal of IKEEE
• /
• v.10 no.1 s.18
• /
• pp.75-86
• /
• 2006

The MANET that can make autonomous distributed Network with Routing function has many differences than past wireless communication. For upcoming ALL-IP environment, MANET device should be connected with wired Internet Network and MANET is required to have a gateway to bridge two different networks to share information from any place. In this paper, Using the GMAHN Algorithm proposed Proactive, Reactive, Hybrid method that provides Inteface between Wired Internet network and MANET, we learned each method's the advantage and disadvantage through the various network environments. And also, we presented the optimization method of Hybrid combined Proactive with Reactive.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.19 no.10
• /
• pp.2367-2372
• /
• 2015

One of various low-density parity-check(LDPC) codes that has been adopted in many communication standards due to its error correction ability is a quasi-cyclic LDPC(QC-LDPC) code, which leads to comparable decoder complexity. One of the main blocks in the QC-LCDC code decoder is a multi-size circular shifter(MSCS) that can perform various size rotation. The MSCS can be implemented with many structures, one of which is based on Banes network. The Benes network structure can perform the normal MSCS operation efficiently, but it cannot use the properties coming from specifications like rotation sizes. This paper proposesd a scheme where the Benes network structure can use the rotation size property with the modification of the control signal generation. The proposed scheme is applied to the MSCS of IEEE 802.16e WiMAX QC-LDPC decoder to reduce the number of MUXes and the critical path delay.

• Journal of Biosystems Engineering
• /
• v.42 no.1
• /
• pp.23-43
• /
• 2017

Purpose: As application-specific wireless sensor networks are gaining popularity, this paper discusses the development and field performance of the GHAN, a greenhouse area network system to monitor, control, and access greenhouse microenvironments. GHAN, which is an upgraded system, has many new functions. It is an intelligent wireless sensor and actuator network (WSAN) system for next-generation greenhouses, which enhances the state of the art of greenhouse automation systems and helps growers by providing them valuable information not available otherwise. Apart from providing online spatial and temporal monitoring of the greenhouse microclimate, GHAN has a modified vapor pressure deficit (VPD) fuzzy controller with an adaptive-selective mechanism that provides better control of the greenhouse crop VPD with energy optimization. Using the latest soil-matrix potential sensors, the GHAN system also ascertains when, where, and how much to irrigate and spatially manages the irrigation schedule within the greenhouse grids. Further, given the need to understand the microclimate control dynamics of a greenhouse during the crop season or a specific time, a statistical assessment tool to estimate the degree of optimality and spatial variability is proposed and implemented. Methods: Apart from the development work, the system was field-tested in a commercial greenhouse situated in the region of Punjab, India, under different outside weather conditions for a long period of time. Conclusions: Day results of the greenhouse microclimate control dynamics were recorded and analyzed, and they proved the successful operation of the system in keeping the greenhouse climate optimal and uniform most of the time, with high control performance.