• Title/Summary/Keyword: Top Down Method

Search Result 390, Processing Time 0.026 seconds

Efficient Triphone Clustering Using Monophone Distance (모노폰 거리를 이용한 트라이폰 클러스터링 방법 연구)

  • Bang Kyu-Seop;Yook Dong-Suk
    • Proceedings of the KSPS conference
    • /
    • 2006.05a
    • /
    • pp.41-44
    • /
    • 2006
  • The purpose of state tying is to reduce the number of models and to use relatively reliable output probability distributions. There are two approaches: one is top down clustering and the other is bottom up clustering. For seen data, the performance of bottom up approach is better than that of top down approach. In this paper, we propose a new clustering technique that can enhance the undertrained triphone clustering performance. The basic idea is to tie unreliable triphones before clustering. An unreliable triphone is the one that appears in the training data too infrequently to train the model accurately. We propose to use monophone distance to preprocess these unreliable triphones. It has been shown in a pilot experiment that the proposed method reduces the error rate significantly.

  • PDF

Estimating Non-Ideal Effects within a Top-Down Methodology for the Design of Continuous-Time Delta-Sigma Modulators

  • Na, Seung-in;Kim, Susie;Yang, Youngtae;Kim, Suhwan
    • JSTS:Journal of Semiconductor Technology and Science
    • /
    • v.16 no.3
    • /
    • pp.319-329
    • /
    • 2016
  • High-level design aids are mandatory for design of a continuous-time delta-sigma modulator (CTDSM). This paper proposes a top-down methodology design to generate a noise transfer function (NTF) which is compensated for excess loop delay (ELD). This method is applicable to low pass loop-filter topologies. Non-ideal effects including ELD, integrator scaling issue, finite op-amp performance, clock jitter and DAC inaccuracies are explicitly represented in a behavioral simulation of a CTDSM. Mathematical modeling using MATLAB is supplemented with circuit-level simulation using Verilog-A blocks. Behavioral simulation and circuit-level simulation using Verilog-A blocks are used to validate our approach.

An experimental study on the tensile performance of joints depending on joint type (접합부 형태에 따른 접합부 인장성능에 관한 실험적 연구)

  • Seo, Hyun-Jae;Park, Jong-sun;Choi, Eun-kyu;Song, Jae-Young;Oh, Sang-Keun
    • Proceedings of the Korean Institute of Building Construction Conference
    • /
    • 2019.05a
    • /
    • pp.190-191
    • /
    • 2019
  • In complex waterproofing method, the stability of the joint is an important factor. The joints should not be broken for repeated movements and tensile forces. Therefore, we tried to confirm the stability of joints in newly developed joints and general joints. Repeated fatigue resistance performance tests and joint tensile strength tests were carried out on four types of joints. The fatigue resistance performance test showed no abnormality in all four types. It was confirmed that the newly developed top-down joint has higher stiffness than existing joints.

  • PDF

A Bottom-up and Top-down Based Disparity Computation

  • Kim, Jung-Gu;hong Jeong
    • Journal of Electrical Engineering and information Science
    • /
    • v.3 no.2
    • /
    • pp.211-221
    • /
    • 1998
  • It is becoming apparent that stereo matching algorithms need much information from high level cognitive processes. Otherwise, conventional algorithms based on bottom-up control alone are susceptible to local minima. We introduce a system that consists of two levels. A lower level, using a usual matching method, is based upon the local neighborhood and a second level, that can integrate the partial information, is aimed at contextual matching. Conceptually, the introduction of bottom-up and top-down feedback loop to the usual matching algorithm improves the overall performance. For this purpose, we model the image attributes using a Markov random field (MRF) and thereupon derive a maximum a posteriori (MAP) estimate. The energy equation, corresponding to the estimate, efficiently represents the natural constraints such as occlusion and the partial informations from the other levels. In addition to recognition, we derive a training method that can determine the system informations from the other levels. In addition to recognition, we derive a training method that can determine the system parameters automatically. As an experiment, we test the algorithms using random dot stereograms (RDS) as well as natural scenes. It is proven that the overall recognition error is drastically reduced by the introduction of contextual matching.

  • PDF

지면구성에 있어서 Proportion에 관한 연구

  • 노상철
    • Journal of the Korean Graphic Arts Communication Society
    • /
    • v.5 no.1
    • /
    • pp.1-10
    • /
    • 1987
  • This paper proposes a design for the software of Graphic system for CAD/CAM, Particularly, 2D Graphic system is studied.The implemention has been tested by GSS PC24 computer and GW-BASIC is used as program language.The system is designed with basic modules of CAD using Top-down method.

  • PDF

AI-Based Object Recognition Research for Augmented Reality Character Implementation (증강현실 캐릭터 구현을 위한 AI기반 객체인식 연구)

  • Seok-Hwan Lee;Jung-Keum Lee;Hyun Sim
    • The Journal of the Korea institute of electronic communication sciences
    • /
    • v.18 no.6
    • /
    • pp.1321-1330
    • /
    • 2023
  • This study attempts to address the problem of 3D pose estimation for multiple human objects through a single image generated during the character development process that can be used in augmented reality. In the existing top-down method, all objects in the image are first detected, and then each is reconstructed independently. The problem is that inconsistent results may occur due to overlap or depth order mismatch between the reconstructed objects. The goal of this study is to solve these problems and develop a single network that provides consistent 3D reconstruction of all humans in a scene. Integrating a human body model based on the SMPL parametric system into a top-down framework became an important choice. Through this, two types of collision loss based on distance field and loss that considers depth order were introduced. The first loss prevents overlap between reconstructed people, and the second loss adjusts the depth ordering of people to render occlusion inference and annotated instance segmentation consistently. This method allows depth information to be provided to the network without explicit 3D annotation of the image. Experimental results show that this study's methodology performs better than existing methods on standard 3D pose benchmarks, and the proposed losses enable more consistent reconstruction from natural images.

A Study on a Recombination Method for the Bottom-up Construction of Spatial Information Products (재조합을 위한 Bottom-up 공간정보제품 제작 방법)

  • Choi, Jae-Yeon;Kim, Eun-Hyung
    • Journal of Cadastre & Land InformatiX
    • /
    • v.47 no.2
    • /
    • pp.185-199
    • /
    • 2017
  • This study is on a recombination method for the construction of spatial information products which demands are unpredictably various. The present production method of digital maps is not flexible enough for their reusability because it is not object-oriented but top-down. Each spatial object needs to have particular attributes to be recombined. The demand changes the production method through the reclassification of data and changing the properties. In a user perspective, the bottom-up method can produce on-demand spatial information products including existing digital maps. The method is derived from case studies and theoretical reviews and compared with the existing production method. In the method spatial information products are reclassified by their geometry objects such as point, line, and polygon, with basic attributes, and other related domain attributes. The geometry objects and domain attributes are connected by adding new attributes for their later relationship and management, which make the recombination possible. To prove its usability of the method it is tested for current and future user demands including the national base map, thematic maps and the future spatial information products.

Synthesis of Nano-Sized Y3Al5O12:Ce3+ Phosphors Prepared by High Energy Beads Milling Process and Their Luminescence Properties

  • Song, Hee-Jo;Kim, Dong-Hoe;Park, Jong-Hoon;Han, Byung-Suh;Hong, Kug-Sun
    • Proceedings of the Korean Vacuum Society Conference
    • /
    • 2012.08a
    • /
    • pp.386-386
    • /
    • 2012
  • For white light emitting diode (LED) applications, it has been reported that Y3Al5O12:Ce3+ (YAG:Ce) in nano-sized phosphor performs better than it does in micro-sized particles. This is because nano-sized YAG:Ce can reduce internal light scattering when coated onto a blue LED surface. Recently, there have been many reports on the synthesis of nano-sized YAG particles using bottom-up method, such as co-precipitation method, sol-gel process, hydrothermal method, solvothermal method, and glycothermal method. However, there has been no report using top-down method. Top-down method has advantages than bottom-up method, such as large scale production and easy control of doping concentration and particle size. Therefore, in this study, nano-sized YAG:Ce phosphors were synthesized by a high energy beads milling process with varying beads size, milling time and milling steps. The beads milling process was performed by Laboratory Mill MINICER with ZrO2 beads. The phase identity and morphology of nano-sized YAG:Ce were characterized by X-ray powder diffraction (XRD) and field-emission scanning electron microscopy (FESEM), respectively. By controlling beads size, milling time and milling steps, we synthesized a size-tunable and uniform nano-sized YAG:Ce phosphors which average diameters were 100, 85 and 40 nm, respectively. After milling, there was no impurity and all of the peaks were in good agreement with YAG (JCPDS No. 33-0040). Luminescence and quantum efficiency (QE) of nano-sized YAG:Ce phosphors were measured by fluorescence spectrometer and QE measuring instrument, respectively. The synthesized YAG:Ce absorbed light efficiently in the visible region of 400-500 nm, and showed single broadband emission peaked at 550 nm with 50% of QE. As a result, by considering above results, high energy beads milling process could be a facile and reproducible synthesis method for nano-sized YAG:Ce phosphors.

  • PDF

Rapid Design Method and System Development for Aircraft Wing Structure

  • Tang, Jiapeng;Han, Jing;Luo, Mingqiang
    • International Journal of Aeronautical and Space Sciences
    • /
    • v.17 no.1
    • /
    • pp.45-53
    • /
    • 2016
  • This work is mainly done by too many manual operations in the aircraft structure design process resulting in heavy workload, low efficiency and quality, non-standardized processes and procedures. A top-down associated design method employing the template parametric technology is proposed here in order to improve the quality of design and efficiency of aircraft wing structure at the preliminary design stage. The appropriate parametric tool is chosen and the rapid design system of knowledge-driven aircraft wing structure is developed. First, a skeleton model of aircraft wing structure is rapidly built up through the template encapsulated design knowledge. Associated design is then introduced to realize the association between the typical structural part and skeleton model. Finally, the related elements are referenced from skeleton model, and a typical structural part reflecting an automatic response for design changes of the upstream skeleton model is quickly constructed within the template. The rapid design system proposed and developed in this paper is able to formalize the design standardization of aircraft wing structure and thus the rapid generation of different aircraft wing structure programs and achieve the structural design knowledge reuse as well.

A Semantic Hierarchy of Korean Nouns using the Definitions of Words in a Dictionary (사전 뜻 풀이말에서 구축한 한국어 명사 의미 계층구조)

  • 조평옥;안미정;옥철영;이수동
    • Korean Journal of Cognitive Science
    • /
    • v.10 no.4
    • /
    • pp.1-10
    • /
    • 1999
  • A Semantic Hierarchy of Korean Nouns(SHKN) where Korean nouns are semantically and hierarchically classified is one of the most important things that can provide semantic information concerned with processing the Korean sentences. In this paper. SHKN is constructed in bottom-up method by making use of the definition of a noun in the Korean dictionary. SHKN constructed in this paper is a forest which consist of 43 trees and 2.443 non-terminal nodes and 10.347 terminal nodes. depth of which is 17. The classes of level 1. 2 of SHKN is quite different from the existing structure classified in top-down method. 2 b but the lower classes than level 2 is very objective.

  • PDF