• Annual Conference of KIPS
• /
• 2022.11a
• /
• pp.646-647
• /
• 2022

Multiview stereo (MVS) 3D reconstruction of a scene from images is a fundamental computer vision problem that has been thoroughly researched in recent times. Traditionally, MVS approaches create dense correspondences by constructing regularizations and hand-crafted similarity metrics. Although these techniques have achieved excellent results in the best Lambertian conditions, traditional MVS algorithms still contain a lot of artifacts. Therefore, in this study, we suggest using a transformer network to accelerate the MVS reconstruction. The network is based on a transformer model and can extract dense features with 3D consistency and global context, which are necessary to provide accurate matching for MVS.

• The Transactions of the Korean Institute of Electrical Engineers A
• /
• v.49 no.1
• /
• pp.31-36
• /
• 2000

To enhance the fault diagnosis ability for the dissolved gas analysis(DGA) of the power transformer, this paper proposes a learning vector quantization(LVQ) for the incipient fault recognition. LVQ is suitable expecially for pattern recognition such as fault diagnosis of power transformer using DGA because it improves the performance of Kohonen neural network by placing emphasis on the classification around the decision boundary. The capabilities of the proposed diagnosis system for the transformer DGA decision support have been extensively verified through the practical test data collected from Korea Electrical Power Corporation.

• KIEE International Transactions on Electrophysics and Applications
• /
• v.11C no.4
• /
• pp.127-132
• /
• 2001

We describe the pulse forming of pulsed $CO_2$laser using multi-pulse superposition technique. A various pulse length, high duty cycle pulse forming network(PFN) is constructed by time sequence. That is, this study shows a technology that makes it possible to make various pulse shapes by turning on SCRs of three PFN modules consecutively at a desirable delay time with the aid of PIC one-chip microprocessor. The power supply for this experiment consists of three PFN modules. Each PFN module uses a capacitor, a pulse forming inductor, a SCR, a High voltage pulse transformer, and a bridge rectifier on each transformer secondary. The PFN modules operate at low voltage and drive the primary of HV pulse transformer. The secondary of the transformer has a full-wave rectifier, which passes the pulse energy to the load in a continuous sequence. We investigated laser pulse shape and duration as various trigger time intervals of SCRs among three PFN modules. As a result, we can obtain laser beam with various pulse shapes and durations from about 250 $mutextrm{s}$ to 600 $mutextrm{s}$.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.22 no.4
• /
• pp.336-344
• /
• 2017

This paper presents the three-port DC-DC converter modeling and controller design procedure, which is part of the solid-state transformer (SST) to interface medium voltage AC grid to bipolar DC distribution network. Due to the high primary side DC link voltage, the proposed converter employs the three-level neutral point clamped (NPC) topology at the primary side and 2-two level half bridge circuits for each DC distribution network. For the proposed converter particular structure, this paper conducts modeling the three winding transformer and the power transfer between each port. A decoupling method is adopted to simplify the power transfer model. The voltage controller design procedure is presented. In addition, the output current sharing controller is employed for current balancing between the parallel-connected secondary output ports. The proposed circuit and controller performance are verified by experimental results using a 30 kW prototype SST system.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.23 no.5
• /
• pp.1-8
• /
• 2023

Cross-domain recommendation is a method that shares related user information data and item data in different domains. It is mainly used in online shopping malls with many users or multimedia service contents, such as YouTube or Netflix. Through K-means clustering, embeddings are created by performing clustering based on user data and ratings. After learning the result through a transformer network, user satisfaction is predicted. Then, items suitable for the user are recommended using a transformer-based recommendation model. Through this study, it was shown through experiments that recommendations can predict cold-start problems at a lesser time cost and increase user satisfaction.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.9
• /
• pp.17-25
• /
• 2023

In this paper, we propose a model that can perform human pose estimation through a MobileViT-based model with fewer parameters and faster estimation. The based model demonstrates lightweight performance through a structure that combines features of convolutional neural networks with features of Vision Transformer. Transformer, which is a major mechanism in this study, has become more influential as its based models perform better than convolutional neural network-based models in the field of computer vision. Similarly, in the field of human pose estimation, Vision Transformer-based ViTPose maintains the best performance in all human pose estimation benchmarks such as COCO, OCHuman, and MPII. However, because Vision Transformer has a heavy model structure with a large number of parameters and requires a relatively large amount of computation, it costs users a lot to train the model. Accordingly, the based model overcame the insufficient Inductive Bias calculation problem, which requires a large amount of computation by Vision Transformer, with Local Representation through a convolutional neural network structure. Finally, the proposed model obtained a mean average precision of 0.694 on the MS COCO benchmark with 3.28 GFLOPs and 9.72 million parameters, which are 1/5 and 1/9 the number compared to ViTPose, respectively.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.11
• /
• pp.81-87
• /
• 2023

Intrusion detection systems that learn metadata of network packets have been proposed recently. However these approaches require time to analyze packets to generate metadata for model learning, and time to pre-process metadata before learning. In addition, models that have learned specific metadata cannot detect intrusion by using original packets flowing into the network as they are. To address the problem, this paper propose a natural language processing-based intrusion detection system that detects intrusions by learning the packet payload as a single sentence without an additional conversion process. To verify the performance of our approach, we utilized the UNSW-NB15 and Transformer models. First, the PCAP files of the dataset were labeled, and then two Transformer (BERT, DistilBERT) models were trained directly in the form of sentences to analyze the detection performance. The experimental results showed that the binary classification accuracy was 99.03% and 99.05%, respectively, which is similar or superior to the detection performance of the techniques proposed in previous studies. Multi-class classification showed better performance with 86.63% and 86.36%, respectively.

• Proceedings of the KIEE Conference
• /
• 2003.07a
• /
• pp.259-261
• /
• 2003

This paper suggests wireless diagnosis and monitoring system using SMS mode of mobile communication network for distribution transformer which could prevent electrical accident in the near future. Data are acquired by measuring the temperature of insulator oil in the distribution transformer and load current. Data acquisition of sensor using mobile communication network carried out filtering of sensor's output to optimize the size of send data Merit of this inspection method is that management, control and monitoring some transformers can be carried out using only one server. This inspection method will be the way of inspection to be worth spotlight in the near future because it is able to solve easily with the minimum facility inspection about state of transformer which is operating, to wide coverage about machine's wrong operation in other field.

• Journal of Electrical Engineering and Technology
• /
• v.12 no.5
• /
• pp.1883-1890
• /
• 2017

Transformers are one of the most precious elements of the electric power system. Stability and reliability of the electric power network mainly depend on the working of the transformer. Leakage reactance of the transformer is one of the important factors and accurate calculation of the leakage reactance is necessary for the transformer designers and electric distributors. Leakage reactance of the transformer depends on the geometry of the transformer. There are many different methods for the calculations of the leakage reactance however mostly are usable when the axial heights of the high voltage and low voltage windings are equal. When the axial heights of high voltage and low voltage windings are asymmetric most of the analytical methods are not reliable. In this study, a new analytical method is introduced for the calculation of the leakage reactance. Fourteen different transformers are investigated in this study and four of them are presented in this paper. The results of the new analytical method are compared with the experimental results. Other analytical and numerical methods are also compared with this new method. Results show that this method is more reliable and accurate as compared to the other analytical methods. The maximum relative error between short-circuit test and proposed method for these fourteen transformers was less than 2.8%.

• Journal of Electrical Engineering and Technology
• /
• v.2 no.3
• /
• pp.300-305
• /
• 2007

This paper reports an investigation of overvoltages caused by ferroresonance in the distribution system, which consists of a bank of open-delta single-phase voltage transformers. The high voltage sides of the voltage transformer are connected to the distribution system via three single-phase fuse cutouts. Due to the saturation characteristic of the transformer cores, ferroresonance can occur in the condition that the transformer is energized or deenergized with single-phase switching operation of the fuse cutouts. The simulation tool based on EMTP is used to investigate the overvoltages at the high side of voltage transformer. Bifurcation diagrams are used to present the ferroresonance behavior in ranges of different operating conditions. The simulation results are in good agreement with the results from the experiment of 22 kV voltage transformers. The mitigation technique with additional damping resistors in the secondary windings of the voltage transformers will be introduced. Brief discussion will be made on the physical phenomena related to the overvoltage and the damage of voltage transformer.