• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.3 no.2
• /
• pp.126-130
• /
• 2002

This paper presents the basic concept of new solid freeform fabrication system using rapid prototyping method. The system could fabricate an arbitrary ceramic form by laser cutting, transferring, stacking, and sintering of each slide. The system consists mainly of laser apparatus. X-Y table, a slide transferring system, and electric oven. The system could fabricate an object that has smooth surface with comparatively short period of time. The system is effective in terms of its direct fabrication capability without second mechanical process. The fabricated shape could directly be used as part of a whole assembly and therefore its method could be applied to various application areas.

• Journal of Information Processing Systems
• /
• v.13 no.5
• /
• pp.1126-1134
• /
• 2017

A variety of medical service applications in the field of the Internet of Things (IoT) are being studied. Segmentation is important to identify meaningful regions in images and is also required in 3D images. Previous methods have been based on gray value and shape. The Visible Korean dataset consists of serially sectioned high-resolution color images. Unlike computed tomography or magnetic resonance images, automatic segmentation of color images is difficult because detecting an object's boundaries in colored images is very difficult compared to grayscale images. Therefore, skilled anatomists usually segment color images manually or semi-automatically. We present an out-of-core 3D segmentation method for large-scale image datasets. Our method can segment significant regions in the coronal and sagittal planes, as well as the axial plane, to produce a 3D image. Our system verifies the result interactively with a multi-planar reconstruction view and a 3D view. Our system can be used to train unskilled anatomists and medical students. It is also possible for a skilled anatomist to segment an image remotely since it is difficult to transfer such large amounts of data.

• Journal of the Korean Institute of Educational Facilities
• /
• v.26 no.4
• /
• pp.19-25
• /
• 2019

Concrete wedge anchor is one of structural components to transfer load of an object attached to a primary structure. Recently, as retrofitting concrete structure is becoming a main issue, mechanical capacity of the anchor should be secured enough. In spite of the structural safety of Cast-in-place anchor, Post-installed anchor is more widely used with ease of placement or change of construction method. However, the post-installed anchors domestically produced have excessive coefficient of variation over 15% of ultimate tensile strength, which yields deteriorated quality in tensile strength. In this research, tensile strength test of anchors, which have improved sleeve and header and produced by a domestic company, was conducted for two variables, concrete strength and effective embedment depth. As a result, enough coefficients of variations were secured in all specimens. Also, in comparison to foreign products, the domestic ones have equal or higher performance.

• International Journal of Advanced Culture Technology
• /
• v.9 no.2
• /
• pp.86-90
• /
• 2021

Sentiment refers to a person's thoughts, opinions, and feelings toward an object. Sentiment analysis is a process of collecting opinions on a specific target and classifying them according to their emotions, and applies to opinion mining that analyzes product reviews and reviews on the web. Companies and users can grasp the opinions of public opinion and come up with a way to do so. Recently, natural language processing models using the Transformer structure have appeared, and Google's BERT is a representative example. Afterwards, various models came out by remodeling the BERT. Among them, the Facebook AI team unveiled the XLM-R (XLM-RoBERTa), an upgraded XLM model. XLM-R solved the data limitation and the curse of multilinguality by training XLM with 2TB or more refined CC (CommonCrawl), not Wikipedia data. This model showed that the multilingual model has similar performance to the single language model when it is trained by adjusting the size of the model and the data required for training. Therefore, in this paper, we study the improvement of Korean sentiment analysis performed using a pre-trained XLM-R model that solved curse of multilinguality and improved performance.

• The Journal of Korea Robotics Society
• /
• v.16 no.1
• /
• pp.56-63
• /
• 2021

As the use of robots in service area increases, research has been conducted to replace human tasks in daily life with robots. Among them, this study focuses on the tidy-up task on a desk using a robot arm. The order in which tidy-up motions are carried out has a great impact on the success rate of the task. Therefore, in this study, a neural network-based method for determining the priority of the tidy-up motions from the input image is proposed. Reinforcement learning, which shows good performance in the sequential decision-making process, is used to train such a task planner. The training process is conducted in a virtual tidy-up environment that is configured the same as the actual tidy-up environment. To transfer the learning results in the virtual environment to the actual environment, the input image is preprocessed into a segmented image. In addition, the use of a neural network that excludes unnecessary tidy-up motions from the priority during the tidy-up operation increases the success rate of the task planner. Experiments were conducted in the real world to verify the proposed task planning method.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.34 no.1
• /
• pp.27-32
• /
• 2021

In this study, we fabricated multilayer graphene on a glass substrate by stacking the monolayer graphene synthesized via chemical vapor deposition. The electrical sheet resistance and optical transmittance were evaluated to confirm the quality of the stacked multilayer graphene. Using the fabricated multilayer graphene/glass structure, we characterized its thermal radiative property in terms of the integrated emissivity. The integrated emissivity of the multilayer graphene/glass structure was tuned from 0.91 to 0.72 when the number of graphene layers was changed from 1 to 12. We also demonstrated that the emissivity tunability provided a way to control the apparent temperature of an object that can be used in infrared stealth applications.

• Journal of The Korean Astronomical Society
• /
• v.55 no.6
• /
• pp.195-205
• /
• 2022

We investigate infrared properties of OGLE4 Mira variables in our Galaxy. For each object, we cross-identify the AllWISE, 2MASS, Gaia, and IRAS counterparts. We present various IR two-color diagrams (2CDs) and period-magnitude and period-color relations for the Mira variables. Generally, the Mira variables with longer periods are brighter in the IR fluxes and redder in the IR colors. In this work, we also revise and update the previous catalog of AGB stars in our Galaxy using the new sample of OGLE4 Mira variables. Now, we present a new catalog of 74,093 (64,609 O-rich and 9,484 C-rich) AGB stars in our Galaxy. A group of 23,314 (19,196 O-rich and 4,118 C-rich) AGB stars are identified based on the IRAS PSC and another group of 50,779 (45,413 O-rich and 5,366 C-rich) AGB stars are identified based on the AllWISE source catalog. For all of the AGB stars, we cross-identify the IRAS, AKARI, MSX, AllWISE, 2MASS, OGLE4, Gaia, and AAVSO counterparts and present various infrared 2CDs. Comparing the observations with the theory, we find that basic theoretical dust shell models can account for the IR observations fairly well for most of the AGB stars.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.17 no.4
• /
• pp.217-228
• /
• 2022

Currently, there are increasing demands for applying deep neural networks (DNNs) in the embedded domain such as classification and object detection. The DNN processing in embedded domain often requires custom hardware such as NPU for acceleration due to the constraints in power, performance, and area. Processing DNN models requires a large amount of data, and its seamless transfer to NPU is crucial for performance. In this paper, we developed a cycle-accurate NPU simulator to evaluate diverse NPU microarchitectures. In addition, we propose a novel technique for reducing the number of memory accesses when processing convolutional layers in convolutional neural networks (CNNs) on the NPU. The main idea is to reuse data with memory interleaving, which recycles the overlapping data between previous and current input windows. Data memory interleaving makes it possible to quickly read consecutive data in unaligned locations. We implemented the proposed technique to the cycle-accurate NPU simulator and measured the performance with LeNet-5, VGGNet-16, and ResNet-50. The experiment shows up to 2.08x speedup in processing one convolutional layer, compared to the baseline.

• Journal of Korea Society of Industrial Information Systems
• /
• v.15 no.5
• /
• pp.149-158
• /
• 2010

A messenger program in this paper was designed and implemented using Use Case Diagram, Class Diagram, Activity Diagram, and Sequence Diagram. We selected total 7 major functionalities such as PC server, login, friends management, chatting, file transfer, account management, etc. The followings are various implementation results from this paper. First, developers could easily understand the desired product and programming flow from using UML Diagrams. Second, PDA users could communicate with developers for a long time without time constraints. Third, there was high usability of UML Diagram in the phase of objected-oriented analysis. Forth, it was easy to update and maintain a program.

• Annual Conference of KIPS
• /
• 2022.11a
• /
• pp.815-817
• /
• 2022

This paper is about the development of an automatic stair climbing trolley for carrying loads without manpower. The design of tri-wheeled structure and center of mass enable the trolley to move on flat ground and also to ascend stairs by self-balancing. The overall design enables the trolley to avoid collision to walls when the trolley rotates on domestic landings. When the camera recognizes the stair, the sensor measures distance from the trolley to the stair. Then the trolley can move to align itself in the middle of the stair and it starts climbing. It can ascend to a specific floor based on the floor number entered by the user. As a result, the automatic stair climbing trolley is expected to help humans by protecting from accidents of dropping loads and saving their power. It is also expected to use for various purposes such as delivering packages, moving and carrying heavy loads in buildings without elevator.