• Proceedings of the KIEE Conference
• /
• 2005.05a
• /
• pp.179-181
• /
• 2005

There is asymmetric in horizontal and vertical side of PDP cell. Every vertical line has BM(Black Mask) to improve luminance contrast. When error diffusion is processed in PDP system, these problems make an error bigger. In 4 inch PDP system, every red, green, blue color of test pattern is presented and each luminance is measured. That is called horizontal(H), diagonal right(R), diagonal left(L) and vertical(V). In red channel, high luminance descending order is V-H-R-L. In green channel, V-H-L-R. In blue channel, V-M-R=L. After average luminance of each direction is calculated. new weighted error diffusion(Weighted ED) is proposed. In digital image signal processing, the error in weighted ED is differ from ED's. The image of weighted ED is more less error compare to conventional ED and close to original image. As the gray level linearity and big size panel is adopted, weighted ED could produce good image.

• Journal of Korean Society of Transportation
• /
• v.13 no.4
• /
• pp.79-103
• /
• 1995

The elasticity of a model is considered most important. Without showing the proper elasticity any model cannot provide useful information for decision making. This paper demonstrates a macro model which can generate dynamic transport informations every 15 minutes. Without the Wardrop principles and the monotonicity assumptions for the link travel time and link volume relationship, the basic elements of this new modeling approache are composed of link density simulation, stochastic incremental route choice, departure time choice, destination choice and mode choice. The elasticity of the proposed model is examined based on elasticity equations and simulation results. Also the transferability from a mega city like Seoul to a big city like Daejon is demonstrated for the choice model. The issues centering around the dynamic relations among density(k), speed(u), and flow rate(v) are also discussed for the modeling of highly congested situations.

• Electronics and Telecommunications Trends
• /
• v.33 no.4
• /
• pp.1-13
• /
• 2018

Deep learning (DL) models have been widely applied to AI applications such image recognition and language translation with big data. Recently, DL models have becomes larger and more complicated, and have merged together. For the accelerated training of a large-scale deep learning model, model parallelism that partitions the model parameters for non-shared parallel access and updates across multiple machines was provided by a few distributed deep learning frameworks. Model parallelism as a training acceleration method, however, is not as commonly used as data parallelism owing to the difficulty of efficient model parallelism. This paper provides a comprehensive survey of the state of the art in model parallelism by comparing the implementation technologies in several deep learning frameworks that support model parallelism, and suggests a future research directions for improving model parallelism technology.

• Electronics and Telecommunications Trends
• /
• v.33 no.4
• /
• pp.43-53
• /
• 2018

As face recognition (FR) has been well studied over the past decades, FR technology has been applied to many real-world applications such as surveillance and biometric systems. However, in the real-world scenarios, FR performances have been known to be significantly degraded owing to variations in face images, such as the pose, illumination, and low-resolution. Recently, visual intelligence technology has been rapidly growing owing to advances in deep learning, which has also improved the FR performance. Furthermore, the FR performance based on deep learning has been reported to surpass the performance level of human perception. In this article, we discuss deep-learning based high-performance FR technologies in terms of representative deep-learning based FR architectures and recent FR algorithms robust to face image variations (i.e., pose-robust FR, illumination-robust FR, and video FR). In addition, we investigate big face image datasets widely adopted for performance evaluations of the most recent deep-learning based FR algorithms.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.28B no.7
• /
• pp.520-534
• /
• 1991

This paper describes the design of an ALU and a shifter for RISC. The RISC datapath is designed to have a 4-stage pipeline and a 20 MHz operating frequency. The ALU makes use of the 32-bit BLC adder which has the characteristics of high speed ane regular structuer and executes the arithmetic instructions-addition and subtraction- and the logical instructions-AND, OR, and XOR. Additionally, multiplication is possible by iterative executions of step instructions to perform shift and add operations. The shifter is implemented by using the modified of funnel shifter. The shifter is able to perform the arithmetic andlogical shift instructions without maskiog. Moreover, it carries out data align operation which conforms to big endian byte address. The logical operation of the desinged ALU and the shifter were simulated using YSLOG and VLSIsim. SPICE simulation results using 1.2um double metal process parameters show that the ALU and shifter have a delay time of 15.9NS and 9.9NS, respectively. Therefore, the ALU and the shifter operates correctly above 20[ MHz ] click ferquency and are composed of about 7K and 15K teansistors, respectively.

• Journal of Computational Design and Engineering
• /
• v.1 no.4
• /
• pp.289-297
• /
• 2014

Although big-sized markers are good for accurate marker recognition and tracking, they are easily occluded by other objects and deteriorate natural visualization and level of immersion during user interaction in AR environments. In this paper, we propose an approach to exploiting the use of rectangular markers to support tangible AR interaction based on fingertip touch using small-sized markers. It basically adjusts the length, width, and interior area of rectangular markers to make them more suitably fit to longish objects like fingers. It also utilizes convex polygons to resolve the partial occlusion of a marker and properly enlarges the pattern area of a marker while adjusting its size without deteriorating the quality of marker detection. We obtained encouraging results from users that the approach can provide better natural visualization and higher level of immersion, and be accurate and tangible enough to support a pseudo feeling of touching virtual products with human hands or fingertips during design evaluation of digital handheld products.

• 전자공학회논문지 IE
• /
• v.48 no.4
• /
• pp.19-24
• /
• 2011

Advances in mobile communications, portable computing and personal electronics have led consumers to demand large viewing area, bright, high resolution, compact and long life portable displays. Pico projectors are the latest technology to prove that big things often do come in small packages. Our goal in this paper was to implement a portable, low power, and highly adaptable miniature pico projector system. The pico projector is designed many functions of the pico projector solutions, multimedia, and T-DMB broadcasting.

• Journal of Korean Society for Quality Management
• /
• v.36 no.4
• /
• pp.37-46
• /
• 2008

There are big differences in six sigma projects between manufacturing processes and transactional/service processes. This paper analyzed the differences between the two areas by examining 18 six sigma case studies in Korean companies. To characterize six sigma case studies, step-by-step questions and check criteria were developed based on 12 step DMAIC methodology. On the basis of those characterized data, the differences between two areas were analyzed. Significant differences were found in the number of CTQs, validation of measurement systems, methods to determine improvement objectives, the methods to elicit improvement plans and types of control systems. The most significant difference was that statistical tools were widely used in manufacturing processes, but qualitative tools were used in transactional/service processes during improvement phase.

• Journal of Environmental Health Sciences
• /
• v.29 no.4
• /
• pp.48-54
• /
• 2003

This study investigated retrogradation properties of Jeungpyun substituted for cereals with dietary fibers of 60% of brown rice, and barley. Quality changes during storage periods of the functional Jeungpyun, were studied using $\alpha$-amylase iodine enzyme digestion methods, X-ray diffraction patterns, and differential scanning calorimetry (DSC). The barley substituted samples showed significantly lower retrogrdation rates than those of control when examined by $\alpha$-amylase method. The Relative crystallity by X-ray diffraction patterns had typical A type in all samples with appealing big crystallity around its diffraction angle 23$^{\circ}$degrees as storage periods were increased. The brown rice and barley Jeunpyun made smaller crystallity than that of control, representing slower retrogradation rates. The batter controls had significantly lower ΔH than the other compared samples when measured by DSC, but had significantly higher ΔH after 30 days of storage, implying that the control required more energy for regelatinization after the 30 days of storage.

• Proceedings of the Korea Concrete Institute Conference
• /
• 1999.04a
• /
• pp.727-732
• /
• 1999

In this research, we made an experiment on the 10 specimen beams that we made. The specimen beams consist of 4 steel plate strengthening beams and 5 carbon fiber sheet strengthening beams. We applied the methods of notch, rounding off a edge, anchor bolt and side shear strengening to the steel plate and for the case of carbon fiber sheet, we applied the methods of anchor bolt, line anchor and shear strengthening. After all the cases were applied, the beams was measured and analyzed about the behavior property of strengthened beams, th ability of strengthening method, the relation between load and the shape of failure, the crack load, the yield load, the shape of crack pattern, the increasing rate from yield load and maximum load and the strain of rebar. All the strengthening methods resulted in almost same value until the yield load, and it wasn't quite different from the theoretical value. In comparison with existing method, the SER, SEAS for the steel plate and the CEA, CESS, CCESS for carbon fiber sheet showed the increasement of ductility with big displacement.