• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.05a
• /
• pp.472-474
• /
• 2019

Because of the difference in network structure and loss function, Verification and identification models have their respective advantages and limitations for person reidentification (re-ID). In this work, we propose a multi-task network simultaneously computes the identification loss and verification loss for person reidentification. Given a pair of images as network input, the multi-task network simultaneously outputs the identities of the two images and whether the images belong to the same identity. In experiments, we analyze the major factors affect the accuracy of person reidentification. To address the occlusion problem and improve the generalization ability of reID models, we use the Random Erasing Augmentation (REA) method to preprocess the images. The method can be easily applied to different pre-trained networks, such as ResNet and VGG. The experimental results on the Market1501 datasets show significant and consistent improvements over the state-of-the-art methods.

• Journal of the Korean Institute of Landscape Architecture
• /
• v.24 no.3
• /
• pp.115-132
• /
• 1996

The purpose of this study is to estimate the soil loss amount with Geographic Information System according to the land use change of Buju mountain area in Mokpo city. To estimate the soil loss, Universal Soil Loss Equation which is the most proper technique to predict soil loss in this site condition is adopted and IDRISI, a raster GIS software, is used. GIS application with USLE is very efficient to estimate soil loss accurately and fastly. In order to decide value and to find application method of USLE factors, we used existing rainfall erosion index, soil erodibility analysis, slope length, slope steepness, vegetation management and practices, which are rated by GIS through the analysis of various studies related USLE. The result of this study was compared with the previous other researches to verify our method of constructing numerical data of USLE's factors. The result of verification of our way showed significance for the soil loss in forest area. But the result of verification for the soil loss in forest area. But the result of verification for the soil loss of cultivated area showed some errors. It seems that this result was due to local variation of topographical map.

• Journal of Environmental Impact Assessment
• /
• v.17 no.4
• /
• pp.243-253
• /
• 2008

The purpose of this study was to simulate the reduction effect of soil loss in the Yongdam reservoir watershed using SWAT model. To evaluate accuracy for flow and sediment yield of SWAT model, calibration was performed for the period from Jan. 2002 to Dec. 2003, and the verification for Jan. 2005 to Dec. 2005. The calibration and the verification were carried out using data observed at the Cheoncheon gaging station. The $R^2$ and EI values in terms of a flow were 0.8 and 0.78 respectively for calibration, whereas they for verification were 0.88 and 0.86 respectively. In terms of a sediment yield, they were 0.7 and 0.48 respectively for calibration, whereas for verification were 0.64 and 0.54 respectively. As a results from model simulation, annual mean soil loss rates in terms of forest, paddy and upland were 0.02 ton/ha/yr, 0.15 ton/ha/yr and 7.58 ton/ha/yr, respectively. The results show that the land use type of a upland has more significant impact on a total soil loss as well as a sediment yield than other types of land use. The sediment delivery ratio was determined to be about 0.35. In this study 2 land cover change scenarios for upland area were considered. These scenarios were used an input to SWAT model in order to evaluate their impact on soil loss and sediment delivery. The results show that a reduction of the upland area would reduce the soil loss and sediment yield.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.05a
• /
• pp.451-454
• /
• 2019

In this paper, we propose a verifiable image transformation networks to transform face sketch to photo and vice versa. Face sketch-photo is very popular in computer vision applications. It has been used in some specific official departments such as law enforcement and digital entertainment. There are several existing face sketch-photo synthesizing methods that use feed-forward convolution neural networks; however, it is hard to assure whether the results of the methods are well mapped by depending only on loss values or accuracy results alone. In our approach, we use two Resnet encoder-decoder networks as image transformation networks. One is for sketch-photo and another is for photo-sketch. They depend on each other to verify their output results during training. For example, using photo-sketch transformation networks to verify the photo result of sketch-photo by inputting the result to the photo-sketch transformation networks and find loss between the reversed transformed result with ground-truth sketch. Likely, we can verify the sketch result as well in a reverse way. Our networks contain two loss functions such as sketch-photo loss and photo-sketch loss for the basic transformation stages and the other two-loss functions such as sketch-photo verification loss and photo-sketch verification loss for the verification stages. Our experiment results on CUFS dataset achieve reasonable results compared with the state-of-the-art approaches.

• Journal of Korea Multimedia Society
• /
• v.15 no.8
• /
• pp.1048-1058
• /
• 2012

This paper proposes the DDMPF(Distributed Data Management Protocol using FAT) which prevents data loss and keeps the security of self-organized storages by comprising a client, a storage server, and a verification server in clouding environment. The DDMPF builds a self-organized storage server, solves data loss by decentralizing the partitioned data in it in contrast to the centralized problem and the data loss caused by the storage server problems of existing clouding storages, and improves the efficiency of distributed data management with FAT(File Allocation Table). And, the DDMPF improves the reliability of data by a verification server's verifying the data integrity of a storage server, and strengthens the security in double encryption with a client's private key and the system's master key using EC-DH algorithm. Additionally, the DDMPF limits the number of verification servers and detects the flooding attack by setting the TS(Time Stamp) for a verification request message and the replay attack by using the nonce value generated newly, whenever the verification is requested.

• Journal of the Korean Society for Precision Engineering
• /
• v.32 no.5
• /
• pp.455-461
• /
• 2015

This paper deals with an experimental verification of a temperature-dependent power loss model of a DC/DC converter in severe temperature conditions. The power loss of a DC/DC converter is obtained by summing the losses by the components constituting the converter including switching elements, diodes, inductors, and capacitors. MIL-STD-810F stipulates that any electronic devices must be operable in the temperature ranging from $-50^{\circ}C$ to $70^{\circ}C$. We summarized the temperature-dependent loss models for the converter components. A SEPIC-type converter is designed and built as a target. Using a constant-temperature chamber, a test rig is set up to measure the power loss of the converter. The experimental results confirm the validity of the loss model within 4.5% error. The model can be useful to predict the efficiency of the converter at the operating temperature, and to provide guidelines in order to improve the efficiency.

• Journal of the Korean Society of Safety
• /
• v.32 no.5
• /
• pp.96-102
• /
• 2017

The highest fatal accident ratio was recorded in the construction industry. According to the industrial insurance premium rate & business type example, among the construction industry, the architectural work has the highest fatal and loss time accident ratio. Previous literature has investigated various aspects of accident occurrence and prevention in architectural work. However, those studied were limited in that they only focused on the fatal accident without considering the loss time accident. But non fatal accidents were recorded more than 50 times of fatal accidents. Therefore non fatal accidents must be controlled to lessen industrial accidents. Based on this, the goal of this study was to investigate the nature of the loss time accident and derive the risk index of work type in architectural work. In this study, opinions of safety experts were gathered and the risk index of work type was derived using AHP(Analytic Hierarchy Process). And verification was accomplished by comparing the results of this study with the risk index derived by analysis of accident records. Results showed that the risk index of work type was significantly higher in steel frame work, temporary installation work, earth & foundation work, facilities work, concrete work. And statistical analysis for verification showed that coefficient of Pearson correlation was 0.686 and P-value was 0.001.

• Proceedings of the KIEE Conference
• /
• 2011.07a
• /
• pp.822-823
• /
• 2011

The authors analyzed harmonic current based loss of a high temperature superconducting (HTS) DC model cable. The loss in HTS DC cable is generated due to the variation of magnetic field caused by harmonic current in a HVDC transmission system. The authors designed and fabricated two meters of HTS DC model cable for verification of real loss characteristic. In this paper, the loss characteristics caused by harmonic current in the HTS DC model cable are analyzed using commercial finite element method software package. The loss of the HTS DC cable is much less than the loss of the HTS AC cable but the loss should be considered to decide a proper capacity of cooling system.

• Proceedings of the Korean Nuclear Society Conference
• /
• 1996.05b
• /
• pp.499-504
• /
• 1996

Current Probabilistic Safety Assessment (PSA) techniques are not usually utilized for day-to-day applications in nuclear power plants. The major reason for this anomaly is the complexity of plant models developed for PSA studies and the multitude of resulting fault trees. This impediment can be overcome by the use of simplified plant models. However, oversimplified models usually result in loss of valuable information and therefore. simplification approaches have to be used judiciously in order to achieve accurate and meaningful results. For this reason. development of an appropriate simplification approach must be performed using extreme caution followed with results verification in sequence as well as system levels. If there are no significant differences between the simplified and the original models, the simplified model can be efficiently used in the application of the PSA. This paper presents a methodology for how to develop a suitable simplification technique and the results of its verification for sample systems and sequences. The results show that the utilization of simplified plant models will significantly reduce the number of fault trees with no significant loss of accuracy.

• Proceedings of the KIEE Conference
• /
• 2002.07d
• /
• pp.2307-2309
• /
• 2002

A simulator had been developed and will be used for reliability verification on turbine control programs for boiler feed pump in power plant prior to its actual operation in field. A mathematical model on thermal dynamics pertaining to prime mover steam turbine and pump was realized and included in this simulator. Also, many design and operating data acquired from fields were utilized in order to decide mechanical and thermal dynamic characteristics such as friction loss windage loss and inertia. A user can decide closing or opening velocity of steam stop valves and steam regulation valves. This simulator is able to generate steam pressure, turbine speed, pump power.